Software structure¶

Thermal Calibration¶

This module allows to calibrate thermal images based on the calibration data obtained in the field. During the calibration process, spectral data is linearly correlated with selected field reflectance spectral.

Tc - Ta¶

This module normalizes the temperature raster image according to the air temperature during the image acquisition.

Convert between formats¶

Since LiDAR processing modules make use of SPDLib tools, the first step is to convert the input dataset into SPD files. There are two types of SPD files, non-indexed and indexed. A format translation module has been included in QGIS to this purpose. The module allows the conversion between different formats and it is also used to re-project data.

The module for format conversion is located in the LiDAR submenu of the ThermoLiDAR Toolbox

ThermoLiDAR plug-in supports SPD/UPD, LAS and a wide range of ASCII formats but the current level of support is not intended to encompass all the available formats. For more information about the formats SPDLib supports, please see the Supported File Formats section.

Parameters¶

Interface to convert between different data formats

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

• Index: The location used to index the pulses and points (required):

  • FIRST_RETURN
  • LAST_RETURN

• Input Format: Format of the input file (Default SPD).

  • SPD: SPD input format with or without spatial index
  • ASCII: ASCII input format
  • LAS/LAZ: Both zipped or normal LAS input format
  • LASNP: LAS input without pulse information

• Ouput Format: Format of the output file (Default SPD).

  • SPD: SPD output format
  • UPD: SPD output format without spatial index
  • ASCII: ASCII output format
  • LAS: LAS output format
  • LAZ: Zipped LAS output format

Optional Input Parameters¶

• Binsize: (float) Bin size for SPD file index (Default 1)
• Schema: (string) schema for the format of the ASCII file being imported
• Input Projection: (string) WKT string representing the projection of the input file
• Output Projection: (string) WKT string representing the projection of the output file
• Num. Columns: (integer) Number of columns within a block (Default 0) - Note values greater than 1 result in a non-sequential SPD file.
• Num. Rows: (integer) Number of rows within a block (Default 25)
• Temporal Path: (string) Path where temporary files can be written to.

Output Parameters¶

• Output: The output SPD file

Merge files¶

In some situations it might be convenient to merge various files into a single SPD file. The merging module merges compatible files into a single non-indexed SPD file. It is possible to provide the projection information of the output file and input files if known.

This module allows displaying classes and returns IDs of the input files with list returns IDs and list classes options, respectively. The ignore checks option forces the input files to be merged in case files come from different sources or have different bin sizes.

The module for merging files is located in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface to merge compatible files into a single non-indexed SPD file

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds (accept multiple files separated by comas).

• Index: The location used to index the pulses and points (required):

  • FIRST_RETURN
  • LAST_RETURN

• Input Format: Format of the input file (Default SPD).

  • SPD: SPD input format with or without spatial index
  • ASCII: ASCII input format
  • LAS/LAZ: Both zipped or normal LAS input format
  • LASNP: LAS input without pulse information

Optional Input Parameters¶

• List Returns IDs: (list of files) Lists the return IDs for the files listed (accept multiple files separated by comas).
• List Classes: (list of files) Lists the classes for the files listed (accept multiple files separated by comas).
• Keep Extent: (Yes/No) Use the extent of the input files as the minimum extent of the output file when indexing the file.
• Source ID: (Yes/No) Set source ID for each input file
• Ignore Checks: (Yes/No) Ignore checks between input files to ensure compatibility
• Schema: (string) schema for the format of the ASCII file being imported
• Input Projection: (string) WKT string representing the projection of the input file
• Output Projection: (string) WKT string representing the projection of the output file

Output Parameters¶

• Output: The output SPD file

Split data into tiles¶

LiDAR data is supplied as flight lines or tiles with different shapes and sizes. It is always useful to divide laser data into equally sized square tiles, though. This helps to store, manage and access data easily. Single tiles should meet memory requirements in order to reduce computational times, which determines the maximum size of each file given an average point density. Overlapping zones between tiles help to prevent border errors and guarantee continuous raster models.

ThermoLiDAR has a built-in tool to create tiles given tiles size and the overlap. Output tiles are saved into the output path (this includes path and prefix) and are named as _rowYYcolXX.spd, being YY and XX the number of row and column of the corresponding tile. Tiles definition is stored in an output XML file (output xml) containing their column, row, extent and core extent (tile extent without overlap).

<tiles columns="23" overlap="50" rows="16" xmax="256500" xmin="239343.510" xtilesize="750" ymax="707224.300" ymin="695246.440" ytilesize="750">
   <tile col="1" corexmax="240093.510" corexmin="239343.510" coreymax="695996.440" coreymin="695246.440" file="" row="1" xmax="240143.510" xmin="239293.510" ymax="696046.440" ymin="695196.440"/>
   <tile col="2" corexmax="240843.510" corexmin="240093.510" coreymax="695996.440" coreymin="695246.440" file="" row="1" xmax="240893.510" xmin="240043.510" ymax="696046.440" ymin="695196.440"/>

   ...

   <tile col="22" corexmax="255843.510" corexmin="255093.510" coreymax="707246.440" coreymin="706496.440" file="" row="16" xmax="255893.510" xmin="255043.510" ymax="707296.440" ymin="706446.440"/>
   <tile col="23" corexmax="256593.510" corexmin="255843.510" coreymax="707246.440" coreymin="706496.440" file="" row="16" xmax="256643.510" xmin="255793.510" ymax="707296.440" ymin="706446.440"/>
</tiles>

The module supports to create single tiles (SINGLE option) by supplying row and column, or to generate the complete set of tiles (ALL option).

The module also creates an auxiliary file listing the input LiDAR which can be eventually kept (keep file list) once the module has finished. It might happen that some tiles are empty; in that case those files can be removed enabling the delete tiles option.

The module for tiling LiDAR data is within the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface for tiling a set of SPD files

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds (accept multiple files separated by comas).

Optional Input Parameters¶

• Extract Tiles: Where to extract

  • ALL: Create all tiles
  • SINGLE: Extract an individual tile given its row and column

• Delete Tiles: (Yes/No) If shapefile exists delete it and then run

• Keep File List: (Yes/No) Keep auxiliary file containing a list of the input files to be tiled

• Tile Size: (float) Size (in units of the coordinate system) of the square tiles (Default: 1000)

• Output Path: (string) The output XML file that contains the tiles definition

Output Parameters¶

• Output XML: The output XML file that contains the tiles definition

Remove Noise¶

Many factors may introduce errors in LiDAR point clouds, including water vapour clouds, multipath, poor equipment calibration, or even a flock of birds. In order to avoid further errors and artefacts in final digital models and poor assess of height metrics, those points have to be removed.

This module removes vertical noise from LiDAR datasets by means of three different. Upper and lower absolute thresholds will clip the file to fit these values. Relative threshold will remove, for each bin within a SPD file, points outside the upper and lower values relative to the median height. Whilst global threshold will use the whole SPD file to calculate the median height and remove points relative to it.

The module for removing noise from data is located in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface for tiling a set of SPD files

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

Optional Input Parameters¶

• Global Rel. Upper Threshold: (float) Global relative to median upper threshold for returns which are to be removed
• Global Rel. Lower Threshold: (float) Global relative to median lower threshold for returns which are to be removed
• Relative Upper Threshold: (float) Relative to median upper threshold for returns which are to be removed
• Relative Lower Threshold: (float) Relative to median lower threshold for returns which are to be removed
• Absolute Upper Threshold: (float) Absolute upper threshold for returns which are to be removed
• Absolute Lower Threshold: (float) Absolute lower threshold for returns which are to be removed

Output Parameters¶

• Output: The output SPD file without noise

Classify Ground Returns¶

Two different classification algorithms have been implemented into the plug-in. These algorithms also called filters allow the classification of the LiDAR points, identifying which point belong to the ground. The filters implement the Progressive Morphology (Zhang et al., 2003; [Zhang2003]) and the Multiscale Curvature (Evans and Hudak, 2007; [EvansHudak2007]) methodologies.

Progressive Morphology filter¶

To classify ground returns an implementation of the Progressive Morphology algorithm (PMF) has been provided. The algorithm QGIS interface has only three options to be set. Under most circumstances the default parameters will be fit the purpose and it is recommended to use the simplest parameters configuration given by default.

The class option allows to apply the filter to particular classes (i.e., if ground returns have been already classified but they need tidying up).

The module that implements the PMF algorithm to classify ground is within the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface of the Progressive Morphology Algorithm to classify ground points

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

Optional Input Parameters¶

• Binsize: (float) Bin size for SPD file index (Default 1)
• Class: (integer) Only use points of particular class
• Ground Threshold: (float) Threshold for deviation from identified ground surface for classifying the ground returns (Default 0.3)
• Median Filter: (integer) Size of the median filter (half size i.e., 3x3 is 1) (Default 2)
• No Median: (Yes/No) Do not run a median filter on generated surface
• Max. Elevation: (float) Maximum elevation difference threshold (Default 5)
• Initial Elevation: (float) Initial elevation difference threshold (Default 0.3)
• Slope: (float) Slope parameter related to terrain (Default 0.3)
• Max. Filter: (float) Maximum size of the filter (Default 7)
• Initial Filter: (float) Initial size of the filter (half size i.e., 3x3 is 1) (Default 1)

Output Parameters¶

• Output: The output SPD file containing classification

Multiscale Curvature filter¶

The plug-ins integrates an implementation of the Multiscale Curvature algorithm (MCC). As before, the QGIS interface has only three options to be set: input, output and class argument. Under most circumstances default parameters for the algorithm will be fit for purpose, but be careful that the bin size used within SPD is not too large as the processing will be at this resolution.

The module that implements the MCC algorithm to classify ground is within the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface of the Progressive Morphology Algorithm to classify ground points

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

Optional Input Parameters¶

• Binsize: (float) Bin size for SPD file index (Default 1)
• Class: (integer) Only use points of particular class
• Median: (Yes/No) Use a median filter to smooth the generated raster instead of a (mean) averaging filter.
• Filter Size: (integer) The size of the smoothing filter (half size i.e., 3x3 is 1; Default = 1)
• Num. Points Tps: (integer) The number of points used for the TPS interpolation (Default = 16)
• Max. Radius Tps: (float) Maximum search radius for the TPS interpolation (Default = 20)
• Step Curve Tolerance: (float) Iteration step curvature tolerance parameter (Default = 0.5)
• Min. Curve Tolerance: (float) Minimum curvature tolerance parameter (Default = 0.1)
• Initial Curve Tolerance: (float) Initial curvature tolerance parameter (Default = 1)
• Scale Gaps: (float) Gap between increments in scale (Default = 0.5)
• Num. Scales Below: (integer) The number of scales below the init scale to be used (Default = 1)
• Num. Scales Above: (integer) The number of scales above the init scale to be used (Default = 1)
• Initial Scale: (float) Initial processing scale, this is usually the native resolution of the data.
• Max. Elevation Threshold: (float) Maximum elevation difference threshold (Default 5)
• Initial Elevation Threshold: (float) Initial elevation difference threshold (Default 0.3)
• Slope: (float) Slope parameter related to terrain (Default 0.3)
• Max. Filter: (float) Maximum size of the filter (Default 7)
• Initial Filter: (float) Initial size of the filter (half size i.e., 3x3 is 1) (Default 1)

Output Parameters¶

• Output: The output SPD file containing classification

Normalise heights¶

SPD files supports both elevation corresponding to a vertical datum and an above-ground height for each discrete return. Before data can be used for generating a Canopy Height Model (CHM) or any height related metric, height field has to be populated. This can be done in two ways. The simplest way is to use a DTM of the same resolution as the SPD file bin size (Image option). The disadvantage of using a DTM is that it is if the DTM is not accurate it can introduce some artefacts. Using this method the only parameters are the input files, both LiDAR file and the DTM, and an output file. The raster DTM needs to the same resolution as the SPD grid and it can be any raster format supported by the GDAL library.

The other option is to interpolate a value for each point generating a continuous surface and reducing any artefacts (Interpolate option). The recommended approach for the interpolation is to use the Natural Neighbour method, as demonstrated by Bater and Coops (2009; [BaterCoops2009]).

The module that defines the height above ground is located in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface of the module to define points heights from the ground

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

Optional Input Parameters¶

• Reference Surface: Select the reference surface for normalise heights

  • Interpolation
  • Image

• Elevation: (raster) The input elevation image

• Binsize: (float) Bin size for SPD file index (Default 1)

• Interpolator: Different interpolation methods to choose from

  • Natural Neighbour
  • Nearest Neighbour
  • TIN Plate

Output Parameters¶

• Output: The output SPD file

Interpolation Module¶

The most common products that can be created from a LiDAR dataset are Digital Terrain Models (DTMs), Digital Surface Models (DSMs) and Canopy Height Models (CHMs). To produce those products, it is necessary to interpolate a raster surface from the classified ground returns and top surface points. Create Digital Model within the ThermoLiDAR plug-in permits to generate these products by choosing model option.

A key parameter is the output raster resolution or binsize which needs to be a multiple of the SPD input file spatial index. Different interpolators can be selected with the interpolator option. This module supports many raster formats by means of the GDAL library.

The interpolation module is in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface to interpolate data

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

• MODEL:

  • DTM: Digital Terrain Model
  • MDS: Digital Surface Model
  • CHM: Canopy Height Model

Optional Input Parameters¶

• Binsize: (float) Bin size for SPD file index (Default 1)

• Interpolator: Different interpolation methods to choose from

  • Natural Neighbour
  • Nearest Neighbour
  • TIN Plate

Output Parameters¶

• Output: The raster file containing the interpolated model

Examples¶

Digital Surface Models (DSMs) are easily done by setting model to DSM. The module will perform an interpolation of the elevation information of all the points of the LiDAR file. The result is a surface representing the ground and all the objects attached to it as can be seen in figure below.

Example of visualization in grey scales of a DSM (1m resolution) generated with the ThermoLiDAR plug-in

In case ground returns have been classified, then interpolating elevation information of ground points will generate a Digital Terrain Model (DTM). For this purpose model has to be set to DTM. The output raster represents the bare ground surface (see next figure).

Example of visualization in grey scales of a DTM (1m resolution) generated with the ThermoLiDAR plug-in

In a forest environment, those points not classified as ground are commonly classified as vegetation. To interpolate the height above ground information of vegetation points produces a Canopy Height Model (CHM). In this case, the model option is set to CHM. The result is raster where canopies are perfectly depicted and ground (see figure below).

Example of visualization in grey scales of a CHM (1m resolution) generated with the ThermoLiDAR plug-in

Generate metrics¶

ThermoLiDAR plug-in is able to calculated different metrics at the same time. Metrics can be simple statistical moments, percentiles of point heights, or even count ratios. Mathematical operators can also be applied to either other metrics or operators.

The XML file required by the metrics option has to be defined a priori with a hierarchical list of metrics and operators. Metrics can be defined manually by typing each XML file which allows the user to adequate the XML file to particular purposes (for more information about XML metrics files, see the How to define metrics? document. However, the user can generate the XML file automatically by means of the module Create metrics XML file.

The module supports different output data formats, that is, raster (all GDAL formats) and vector. Raster option extends the output to the entire input file, assessing the metrics for each pixel the final raster output and creating as many bands as metrics have been defined within the XML file. Vector option requires an input shapefile containing polygon entities. The output shapefile database will be populated with the metrics computed inside the polygons.

The module to generate metrics is located in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface to calculate metrics

Required Input Parameters¶

• Input: SPD file that contains the LiDAR point clouds.

• Metrics: (file) XML file containing the metrics template

• Output Data Format:

  • Image: Raster output
  • Vector: Vector output

Required Input Parameters¶

• Binsize: (float) Bin size for processing and the resolution of the output image. Note: 0 will use the native SPD file bin size
• Vector File: (shapefile) Input shapefile (only with vector output).

Output Parameters¶

• Output: The raster file containing the interpolated model

The figure below shows the Height 95th percentile computed for the same dataset than the previous examples:

Height 95th percentile computed into a raster image (10m resolution)

Create metrics XML file¶

The module offers the possibility to create the file with a single metric selected from a list. However, the list also accept the options ALL, FOREST and PERCENTILES:

• ALL: Selects all the metrics listed in the option list.
• FOREST: Includes some metrics that are commonly used in forest applications. This metrics includes percentiles from 99th to 50th, groundCover, canopyCover, maxHeight and meanHeight.
• PERCENTILES: Includes all percentiles from 99th to 10th.

In case other metrics than those available in the list are needed, the user can specify them by writing their names separated by ; (semi-colon, and without spaces) in the field labelled with string format. This could be an example:

pH99;pH98;pH95;pH90;pH80;maxHeight;canopyCover;NumberReturnsNoGround

In this case, the module will omit any selection made in the first input parameter and it will create the XML file with the metrics supplied by the user. In case the string any metric is misspelled, the module will inform the user of the mistake and omit the metric. The file will be created with the metrics that are correctly written.

The module to generate metrics XML file is located in the LiDAR submenu of the ThermoLiDAR toolbox

Parameters¶

Interface of the module to generate XML file containing metrics

Optional Input Parameters¶

• List of metrics: Metrics to be used
• String of metrics: Metrics to be used

Output Parameters¶

• Output: The output XML file containing metric definitions

References

[Bunting2013]

Bunting, P., Armston, J., Clewley, D., Lucas, R. M., 2013. Sorted pulse data (SPD) library. Part II: A processing framework for LiDAR data from pulsed laser systems in terrestrial environments. Computers and Geosciences 56, 207 – 215.

[BaterCoops2009]

Bater, C. W., Coops, N. C., 2009. Evaluating error associated with lidar-derived DEM interpolation. Computers and Geosciences 35 (2), pp. 289–300.

[EvansHudak2007]

Evans, J. S., Hudak, A. T., 2007. A multi-scale curvature algorithm for classifying discrete return lidar in forested environments. IEEE Transactions on Geoscience and Remote Sensing 45 (4), pp. 1029 – 1038.

[Naesset1997a]

Naesset, E., 1997. Determination of mean tree height of forest stands using airborne laser scanner data. ISPRS Journal of Photogrammetry and Remote Sensing, 52, pp. 49 – 56.

[Naesset1997b]

Naesset, E., 1997. Estimating timber volume of forest stands using airborne laser scanner data. Remote Sensing of Environment, 61, pp. 246 – 253.

[Naesset2002]

Naesset, E., 2002. Predicting forest stands characteristics with airborne scanning laser using a practical two-stage procedure and field data. Remote Sensing of Environment, 80, pp. 88 – 99.

[Zhang2003]

Zhang, K., Chen, S., Whitman, D., Shyu, M., Yan, J., Zhang, C., 2003. A progressive morphological filter for removing nonground measurements from airborne LIDAR data. IEEE Transactions on Geoscience and Remote Sensing 41 (4), pp. 872 – 882.

Forest stands segmentation¶

Within OD tools, users are willing to choose between developing a semi-automatic segmentation and using a pre-defined object feature. Segmentation tools are based on algorithms that segment an image into areas of connected pixels based on the pixel DN value. ThermoLiDAR image segmentation tools will be based on region growing algorithms. The basic approach of a region growing algorithm is to start from a seed region (typically one or more pixels) that are considered to be inside the object to be segmented. The pixels neighbouring this region are evaluated to determine if they should also be considered part of the object. If so, they are added to the region and the process continues as long as new pixels are added to the region. Region growing algorithms vary depending on the criteria used to decide whether a pixel should be included in the region or not, the type connectivity used to determine neighbours, and the strategy used to visit neighbouring pixels. Image segmentation is a crucial step within the object-based remote sensing information retrieval process. As a step prior to classification the quality assessment of the segmentation result is of fundamental significance for the recognition process as well as for choosing the appropriate approach and parameters for a given segmentation task. Alternatively, user could be interested on using a pre-defined object feature. This object feature could be a segmentation shape file provided from other source or any other land cover mapping. Also, the user can use a pre-defined regular object, defining the size of the square to be used previously.

Forest Condition Levels¶

The most critical part in applying forest health condition indicators is the user’s accuracy defining forest degradation levels. Besides user’s training another critical factor is to select under analysis a robust physiological indicator and to carry out an accurate field measurements campaign.

Potential physiological indicators of forest decline such us pigment concentration, photosynthesis, respiration and transpiration rate holds great potential to shed light on the mechanisms and processes that occur as a result of drought stress. In the short-term, climate can change the physiological conditions of the forest resulting in acute damage, but chronic exposure usually results in cumulative effects on physiological process. These factors effects on the plants light reactions or enzymatic functions and increased respiration from reparative activities. Gradual decreases in photosynthesis, stomatal conductance, carbon fixation, water use efficiency, resistance to insect and cold resistance were found in most of trees which are very typical symptom of stress conditions

Long-term exposure of water stress to a combination of high light levels and high temperatures causes a depression of photosynthesis and photosystem II efficiency that is not easily reversed, even for water-stress-resistant forest species. The decrease in the photochemical efficiency of photosystem II (ΦPSII) is related to the conversion of violaxanthin to antheraxanthin and zeaxanthin produced by an increase in harmless non-radiative energy dissipation (qN) and providing photo-protection from oxidative damage. One of the most widely physiological indicator applied in the analysis of long-term effect on forest health condition is de Leaf Area Index (LAI). The following is an example of the statistical analysis performs on LAI values measured from an Oak forest inventoried in the framework of THERMOLIDAR project.

Structurally Homogeneous Forest Units¶

This tool provides the tools required for the classification of forest stands structurally different. The classification is based on two main structural parameters, average height of the trees and density. Input data needed to run this process is obtained from LiDAR data. Alternatively, user can provide external forest maps in a .shp format type with an attribute of the number of class. The following figure shows an example of the units defined for the oak forest under analysis. Using a grey scale, trees were grouped in 3 classes with significant differences in terms of structural composition.

Forest Health Monitoring¶

The main function of this tool is defining health condition differences in the vegetation at the stand level. Input parameters defined by users should mainly contain: thermal imaging data and the FSC Polygons (vector file with structurally homogeneous stands. Forest stands included in this analysis should be carried specifically based on one species. The user can perform a supervised or an unsupervised classification depending of the availability of field data measurements to define training areas. It should be highlight, that at this point of the analysis, users are willing to obtain an integrated mapping of forest health distribution levels based on thermal data but also standardized by forest stands units defined from lidar-based metrics. The following figure, shows an example of the units defined for assessment of the status of forest condition. Using a colour palette, trees were grouped in different classes with significant differences in terms of structural composition and physiological status. The colour palette ranges from red to green, where red colour is relate with trees with high level of damage and green colour represents trees with optimum health condition.

Health Condition Levels¶

Shapiro Test¶

Before proceeding with the classification of items by level of damage according to several variables taken in the field, we verify that the set of physiological variables follow a normal distribution. For this we use the Shapiro test, located in the toolbox [Analysis] Health Condition Level > Shapiro Test.

Required Input Parameters¶

• Input vector: Vector file that contains information on physiological data.
• Parameter: Vector’s field to analyse if it follows the normal distribution

Ouput Parameters¶

• R Console Output: File with the output result. The output is composed of the following values

  • Statistic - The value of the Shapiro-Wilk statistic.
  • p.value - An approximate p-value for the test. This is said in Royston (1995) to be adequate for p.value < 0.1.
  • Method - The character string “Shapiro-Wilk normality test”.
  • data.name - A character string giving the name(s) of the data.

Standardize¶

To be able to use the variables in our analysis, they should follow a normal distribution. One way to force these follow a normal distribution is by definition. This characterization involves the conversion of the variable that follows a distribution N (μ, σ), a new variable with distribution N (1,0). This tool is situated in [Analysis] Heath condition level > Standardize.

Required Input Parameters¶

• Input vector: Shape that contains information on physiological data.
• Variable: Shapefile’s field to standardize

Ouput Parameters¶

• Output vector: The user will output a new shapefile with the standardized variable.

Clustering¶

This tool allows us to group one or more physiological variables according to their degree of similarity between individuals in the sample. So, the goal of clustering is to determine the intrinsic grouping in a set of unlabelled data. But how to decide what constitutes a good clustering? It can be shown that there is no absolute best criterion which would be independent of the final aim of the clustering. Consequently, it is the user which must supply this criterion, in such a way that the result of the clustering will suit their needs.

To make this tool has been chosen by a hierarchical approach. The user-supplied items are categorized into levels and sublevels within a class hierarchy, forming a hierarchical tree structure.

Required Input Parameters¶

• Input vector: Shape containing information on physiological data.
• Cols names: Name the columns containing details physiology (separated by semicolons ‘;’)
• Number of groups: Number of groups

Ouput Parameters¶

• R plots: File with the R output result.
• Output vector: Output shapefile with a new variable (group) indicating that each group is member.

ANOVA¶

ANOVA test are conducted for each variable to indicate how well the variable discriminates between clusters.

The hypothesis is tested in the ANOVA is that the population means (the average of the dependent variable at each level of the independent variable) are equal. If the population means are equal, it means that the groups did not differ in the dependent variable, and consequently, the independent variable is independent of the dependent variable.

Required Input Parameters¶

• Input vector: Shapefile that contains information about the cluster
• Dependent variable: Field shape that acts as a dependent variable
• Independent variable: Field shape that acts as an independent variable

Ouput Parameters¶

• R Console Output: File with the R output result. The result will be a list of ANOVA tables, one for each response (even if there is only one response”. They have columns “Df”, “Sum Sq”, “Mean Sq”, as well as “F value” and “Pr(>F)” if there are non-zero residual degrees of freedom. There is a row for each term in the model, plus one for “Residuals” if there are any.

Structurally Homogeneous Forest Units¶

The analytical purpose of this tool is the definition of structurally homogeneous stands that allow us to minimize the effects of structure on the thermal information, and therefore allow us to obtain related health outcomes woodland. To do this, the software allows the user to define the structure of the stand from the height data. The calculation of uniformity is therefore a function of two variables directly derived from LiDAR data the 95th percentile obtained from MDV and the penetration rate, obtained from density points that penetrate the forest canopy.

• Polygons: Vector file that contains polygons to classify.
• ID: Vector’s field that that indicates the id of each item
• Equation: Operator used to classify the feats.
• Clusters: Number of output groups. The value is 3 by default.
• Cutoff: Stop threshold algorithm. The value is 0.5 by default.

• Output: Vector file name containing the classification.

Forest Health Classification¶

Unsupervised Pixel-based Classification¶

The user has a stratification of the study area, and classified based on the structure. Through this tool, a classification of the pixels of temperatures will be performed, based on the classification of defined structure. Thus, the output will be a raster temperature for each of the groups of homogeneity.

Required Input Parameters¶

• Temperature raster layer: Input temperature raster
• SHFU Polygons: Vector layer that contains structurally homogeneous stands.
• SHFU Field: SHFU layer field containing the group that belong each item.
• Clusters: Number of output classes.
• Cutoff: Stop threshold. The value is 0.5 by default.

Ouput Parameters¶

• Output: Raster file name containing the classification of temperatures based on homogeneous stand units.
• Statistics: CSV File containing statistics for groups.

Unsupervised Object-oriented Classification¶

The user has a stratification of the study area, and classified based on the structure. Through this tool, a classification of the objects of temperatures mean will be performed, based on the classification of defined structure. Thus, the output will be a raster temperature for each of the groups of homogeneity.

Required Input Parameters¶

• Temperature raster layer: Input temperature raster
• SHFU Polygons: Vector layer that contains structurally homogeneous forest units.
• SHFU Field: SHFU layer field containing the group that belongs each item.
• Clusters: Number of output classes.
• Cutoff: Stop threshold. The value is 0.5 by default.

Ouput Parameters¶

• Output: Raster file name containing the classification of temperatures based on homogeneous stand units.
• Statistics: CSV File containing statistics for groups.

Supervised Pixel-based Classification¶

Similarly as in the previous point, the user can perform a classification of the temperature response to the stand of homogeneity. Unlike supervised classification, the user has a number of AOIs that guide the classification process.

Required Input Parameters¶

• Temperature raster layer: Input temperature raster
• SHFU Polygons: Vector layer that contains structurally homogeneous stands.
• SHFU Field: SHFU layer field containing the group that belong each item.
• ROIs vector: Vector file with training areas.
• ROIs - SHFU Identifier: Vector’s field that contains the group’s homogeneity that belongs each item of training areas.
• ROIs - FSC Identifier: Vector’s field that contains the group “Forest health level” that belongs within their group of homogeneity.
• k: Number k nearest neighbours. The value is 3 by default.

Ouput Parameters¶

• Output: Raster file name containing the classification of temperatures based on homogeneous stand units.
• Statistics: CSV File containing statistics for groups.

Supervised Object-oriented Classification¶

Required Input Parameters¶

• Temperature raster layer: Input temperature raster
• SHFU Polygons: Vector layer that contains structurally homogeneous stands.
• SHFU Field: SHFU layer field containing the group that belong each item.
• ROIs vector: Vector file with training areas.
• ROIs - SHFU Identifier: Vector’s field that contains the group’s homogeneity that belongs each item of training areas.
• ROIs - FSC Identifier: Vector’s field that contains the group “Forest health level” that belongs within their group of homogeneity.
• k: Number k nearest neighbours. The value is 3 by default.

Ouput Parameters¶

• Output: Raster file name containing the classification of temperatures based on homogeneous stand units.
• Statistics: CSV File containing statistics for groups.

Stepwise Multivariate Regression Model¶

The Stepwise Multivariate Regression model was firstly introduce by Naesset (1997a, 1997b) to estimates tree heights ([Naesset1997a]) and volumes ([Naesset1997b]). The methodology assumes that stands have been previously classified before sample plots are measured. Plots must have the same size and have to be regularly distributed throughout the study area. Height metrics derived from LiDAR have to be calculated within each single sample plot (see `Generate metrics`_) excluding points lower than 2 meters, so that stones and shrubs are avoided. Metrics have to include (Naesset 2002; [Naesset2002]):

• Quantiles corresponding to the 0th, 5th, 10th, 15th, ..., 90th, 95th, 98th percentiles of the distribution.
• The maximum values
• The mean values
• The coefficients of variation
• Measures of canopy density

Naesset definition of canopy density is considered as the proportions of the first echo laser hits above 10th, ..., 90th percentiles of the first echo height distribution to total number of first echoes.

For each sample plot a logarithmic regression equation is formulated:

\[Y = \beta_0 h_0^{\beta_1} h_{10}^{\beta_2} \ldots h_{90}^{\beta_{11}} h_{max}^{\beta_{12}} h_{mean}^{\beta_{13}} d_{10}^{\beta_{14}} \ldots d_{90}^{\beta_{23}} \ldots\]

which, in linear form is expressed as:

\[\begin{split}\ln Y &= \ln\beta_0 + \beta_1 \ln h_0 + \beta_2 \ln h_{10} \ldots + \beta_{11} \ln h_{90} + \beta_{12} + \\ &+ \ln h_{max} + \beta_{13} \ln h_{mean} + \beta_{14} \ln d_{10} + \ldots + \beta_{23} \ln d_{90}^{\beta_23} + \ldots\end{split}\]

Where \(Y\) are the field values (dependent variable); \(h_{i}\) are height percentiles; \(h_{max}\) and \(h_{mean}\) are maximum and mean height, respectively; and \(d_{j}\) are Naesset canopy densities.

The module to calibrate the the forest model is located in [Analysis] Forest Model submenu of the ThermoLiDAR toolbox

Parameters¶

Interface to calculate calibrate the forest model

Required Input Parameters¶

• Layer: Shapefile layer that contains the LiDAR metrics.
• Dependent Variable: Vector field containing the observations of the predictable variable
• Independent Variable i: Vector field containing the LiDAR metrics that will be use to characterized the model

Output Parameters¶

• Output: (html) File containing the final report of the multivariate regression

Hybrid model¶

This module implements a hybrid model to infer some forestry variables. The equations of the model can be written as:

where \(TH\) is Top Height, \(SI\) is Site Index, \(Yield\) is Yield class, \(Mean_{DBH}\) is Mean DBH, \(Trees_{ha}\) is the number of trees per hectare, \(Volume_{ha}\) is the volume of biomass per hectare, \(Basal\_Area_{ha}\) is the basal area per hectare; and \(\left\{\alpha_1, \alpha_2, \alpha_3, \beta_1, \beta_2, b_1, b_4, c_1, c_4, g_1, g_3, h_1, h_4\right\}\) depends on the specie. \(TH\) is well correlated with the \(95^{th}\) height percentile for Norway Spruce and Sitka Spruce, and with the \(90^{th}\) height percentile for Scot Pine, depending on the specie.

In case no information about trees age is provided, a singularity occurs in the Site Index and Yield Class equations, and thus these variables cannot be then computed.

If the study area contains various stands with different species and ages, then users could be interested in using a layer with sub-compartmentation information. In this case, both age and vegetation layers can be provided in order to overwrite the default constant values of age and vegetation. Layers require to specify the name of the attribute column for each parameter (Age column and Vegetation column).

The output raster will contain 7 different bands, one per each different estimated variable: Top Height, Mean DBH, number of trees per hectare, Volume per hectare, Basal Area per hectare, Site Index and Yield Class. However, when age has been set by default, that is equal to 0, only 5 bands are created in the output raster due to the restrictions in the equations mentioned above.

This figure should depict where to find the module

Parameters¶

This figure should show the module graphical interface

Required Input Parameters¶

• Canopy: (raster) Canopy height layer
• Age: (integer): Age of the vegetation (years) (Default 0)

Optional Input Parameters¶

• Age layer: (vector) Sub-compartmentation at stand level with age layer

• Age column: (string) Name of the column attribute for age sub-compartmentation

• Vegetation layer: (vector) Sub-compartmentation at stand level with species information

• Vegetation column: (string) Name of the column attribute for species sub-compartmentation

• Vegetation type:

  • Sitka Spruce
  • Norway Spruce
  • Scots Pine

Output Parameters¶

• Output: Output raster

Canopy stomatal conductance¶

The module calculates canopy stomatal conductance (TC) of an area based on the information supplied by canopy temperature, terrain, canopy model, LAI, vegetation type and weather conditions amongst others parameters. The model implements the methodology proposed by Blonquist et al., 2009 ([blonquist2009]).

The user is required to provide raster layers for canopy temperature, digital terrain model and canopy height model as input parameters, an output raster layer for the output and the day of the year (DOY). DOY is needed in order to calculate the solar radiation of the particular day and hour of the year when measurements were taken, and must be provided as DDMMYY:hhmm, where DD is the day of the month, MM is the month of the year, YYYY is the year, hh is the hour of the day (24h format) and mm is the minutes of the hour.

LAI, wind speed and air temperature parameters have default values that will be used as constants throughout the whole area. Users can provide different values or even raster layers that to continuously cover the study area. Latitude, precipitation, pressure, humidity should be set to meet location and the weather conditions of the thermal survey. Parameters such as quantum yield efficiency, fraction of photosynthetic active radiation, light extinction coefficient, or atmospheric turbidity factor are optional and default values can be used in most cases.

Ground coverage reflects energy in different ways depending on the coverage type. Vegetation type can be selected from a list in order to set the corresponding energy reflection coefficient. Types available are: * Crop: 0.31 * Deciduous low sun: 0.29 * Deciduous high sun: 0.23 * Conifers: 0.312 * Grass: 0.24 * Snow: 0.85

This figure should depict where to find the module

Parameters¶

This figure should show the module graphical interface

Required Input Parameters¶

• Canopy temperature: (raster) Canopy temperature layer
• Digital terrain model: (raster) Digital terrain model layer
• Canopy hight model: (raster) Canopy hight model layer

Optional Input Parameters¶

• LAI layer: (vector) Leaf area index layer

• LAI: (float) Leaf area index constant value (m2/m2) (Default: 3.0)

• Wind layer: (vector) Wind speed layer

• Wind: (float) Wind speed (m s-1) (Default: 2.0)

• Air Temperature layer Air temperature layer

• Temperature (float) Air temperature constant value (Default: 0.0)

• Vegetation type:

  • Conifers
  • Crop
  • Deciduous low sun
  • Deciduous high sun
  • Grass
  • Snow

• DOY: (string) Day of the year to calculate sun radiance (Format: DDMMYYYY:hhmm)

• Latitude: (float) Approximate latitude of study area (degrees)

• Pressure: (float) Barometric pressure (mb) (Default: 1013.25)

• Precipitation: (float) Precipitation (mm) (Default: 0.0)

• Humidity: (float) Relative humidity (%) (Default: 0.00)

• Wind height: (float) Wind reference height (m) (Default: 0.0)

• Temperature height: (float) Temperature reference height (m) (Default: 0.0)

• Yield: (float) Quantum yield efficiency (mol C mol-1 PAR) (Default: 0.05)

• Active radiation: (float) Fraction of photosynthetic active radiation (Default: 0.45)

• Respiration: (float) Fraction of respiration (Default: 0.5)

• Light extinction coefficient: (float) Light extinction coefficient (Default: 0.5)

• Linke: (float) Atmospheric turbidity factor (Default: 3.0)

Output Parameters¶

• Output: Output raster containing stomatal conductance estimation

References

[blonquist2009]

Blonquist, J. M., Norman, J. M., Bugbee, B., 2009. Automated measurements of canopy stomatal conductance based on infrared temperature. Agricultural and Forest Meteorology, 149, 1931-1945

Convert File Formats - spdtranslate¶

The spdtranslate command is one of the key commands associated with SPDLib as it allows for the conversion between the various supported file formats, while it also supports coordinate system conversion.

Although, the most common use of this tool is converting data to the SPD format. The simplest command for converting to UPD (SPD with a spatial index) is shown below, where the input and output file formats have been specified alongside the field used to attribute each pulse with a location to be used if the pulses where later index (i.e., into an SPD file):

spdtranslate --if LAS --of UPD -x FIRST_RETURN -i QueenElisabeth_example.las -o QueenElisabeth_example.spd

If you wish to explicitly define the projection of the SPD file then use the --input_proj and --output_proj switch to specify a text file containing the OGC WKT string representing the projection. The following command provides an example where the coordinate system (UK Ordnance Survey national grid) has been specified when converting data to an SPD file:

spdtranslate --if LAS --of UPD -x FIRST_RETURN --input_proj ./OSGB1936.wkt -i QueenElisabeth_example.las -o QueenElisabeth_example.spd

While the following command converts from WGS84 to UK Ordnance Survey national grid while reading the file and converting to SPD:

spdtranslate --if LAS --of SPD --convert_proj --input_proj WGS84.wkt --output_proj OSGB1936.wkt -xFIRST_RETURN -b 1 -i QueenElisabeth_example.las -o QueenElisabeth_example.spd

To convert data to the SPD format, where data is indexed on to a 10 m grid the following command is the simplest form.:

spdtranslate --if LAS --of SPD -x FIRST_RETURN -b 10 -i QueenElisabeth_example.las -o QueenElisabeth_example.spd

ThermoLiDAR plug-in supports SPD/UPD, LAS and a wide range of ASCII formats but the current level of support is not intended to encompass all the available formats. For more information about the formats SPDLib supports, please see the Supported File Formats section.

In the examples given above the whole input files are read into memory and sorted into the spatial grid before being written output the file. This requires that you have sufficient memory to store the whole dataset and index data structure in memory. If you do not have sufficient memory you can use the --temppath option. This will tile the file into the disk while building the SPD.

Format convertion can be easily done in QGIS with the Processing Toolbox ‣ [Processing] LiDAR ‣ Convert between formats tool:

Some options are set by default as the format of the input (LAS) and output (SPD) files, and the location used to index pulses (FIRST_RETURN) and the user has to set the other options manually. The two only required options are the input and output file. If the absolute path of the input file is know it can be written directly, otherwise there exists the possibility of search it by clicking the button on the right. This will open another interface.

The output file path is specify in the same way by clicking the button on the right of the output file field. In this case, the user search for the path where the output file should to be saved into and writes the name of the file. It is worth to notice that the file extension has to be set and it should agree with the output format.

Once both files have been set, the graphical interface should look like this

and it should be ready to be executed by clicking OK. To be sure it worked properly, the user can visualise the new spd file with the SPD Points Viewer or load it into QGIS by selecting Processing Toolbox ‣ [Processing] LiDAR ‣ Visualise SPD file:

To load the spd file click on the button and the user will be ask to select a file.

Once the spd file has been loaded, QGIS will show the result in the canvas:

Classify Ground Returns¶

Progressive Morphology Filter - spdpmfgrd¶

The spdpmfgrd command is an implementation of the progressive morphological filter algorithm of [Zhang2003]. The algorithm works by generating an initial minimum return raster surface at the bin resolution of the SPD file. Circulate morphological operators of a range of scales. At each scale a morphological closing (erosion + dilation) operation is performed. The new height value from the morphological operator is kept if it is above the elevation difference threshold where the elevation difference threshold is increased between threshold using a slope value.

Under most circumstances the default parameters for the algorithm will be fit for purpose, but be careful that the bin size used within SPD is not too large as the processing will be at this resolution:

spdpmfgrd -i QueenElisabeth_example.spd -o QueenElisabeth_example_pmfgrd.spd

The QGIS interface of this tool (Processing Toolbox ‣ [Processing] LiDAR ‣ Classify points (Progressive Morphology algorithm)) has only three options that are required. Apart from the input and output files, you can select the class the filter will be applied to.

Filter points depending on class¶

The --class option allows the filter to be applied to returns of a particular class (i.e., if you have ground returns classified but it needs tidying up etc). It’s useful for TLS as it can take a thick slice with mcc algorithm and then use the PMF algorithm to tidy that result up to get a good overall ground classification.

Multi-Curvature Classifier – spdmccgrd¶

The spdmccgrd command is an implementation of the multi-scale curvature algorithm [EvansHudak2007]. This algorithm was created at the US Forest Service and does a good job at classifying ground returns under a forest canopy while retaining the terrain but it does not differentiate the buildings. Under most circumstances the default parameters for the algorithm will be fit for purpose and it is recommend that you try these first with the following command.:

$ spdmccgrd -i QueenElisabeth_example.spd -o QueenElisabeth_example_mccgrd.spd

The QGIS interface of this tool (Processing Toolbox ‣ [Processing] LiDAR ‣ Classify points (Multiscale Curvature Algorithm)) has only three options that are required. Apart from the input and output files, you can select the class the filter will be applied to.

Filter points depending on class¶

As in the spdpmfgrd case, the --class option allows the filter to be applied to returns of a particular class. It’s useful as it can take a thick slice with PMF algorithm and then use the MCC algorithm to tidy that result up to get a good overall ground classification.

Combining filters¶

Another option which can improve the ground return classification is to combine more than one filtering algorithm to take advantage of their particular strengths and weaknesses. A particularly useful combination is to first run the PMF algo- rithm where a ‘thick’ slice is taken (e.g., 1 or 2 metres above the raster surface) and then the MCC is applied to find the ground returns (using the --class 3 option):

spdmccgrd -i --class 3 QueenElisabeth_example_pmfgrd.spd -o QueenElisabeth_example_mccgrd.spd

Normalise Heights - spddefheight¶

The spddefheight command is used to define the height field within both the pulse and point fields of the SPD data file. This can be done in two ways, the simplest is the use of a DTM of the same resolution as the SPD file bin size. The disadvantage of using a DTM is that it is in effect using a series of spot heights and this can introduce artefacts. Therefore, interpolating a value for each point/pulse generates a continuous surface reducing any artefacts. The recommended approach is to use Natural Neighbour interpolation, as demonstrated in the paper of [BaterCoops2009].

Interpolation Mode¶

The interpolators are the same as those defined within the spdinterp command so look to the Interpolate DTM and DSM section for details on their use. The recommend command using the natural neighbour interpolation algorithm, along with the default parameters is:

spddefheight --interp --in NATURAL_NEIGHBOR -i QueenElisabeth_example_mccgrd.spd -o QueenElisabeth_example_height.spd

Within QGIS just select Processing Toolbox ‣ [Processing] LiDAR ‣ Normalise heights:

Interpolate DTM and DSM¶

The most common product to be created from a LiDAR SPD dataset are Digital Terrain Model (DTM), Digital Surface Model (DSM) and Canopy Height Models (CHM). To produce those products you need to interpolate a raster surface from the classified ground returns and top surface points. A key parameter is the resolution of the raster which is generated, within SPDLib the resolution of the raster needs to be a whole number multiple of the SPD index, for example, if the SPD file has a bin size of 10 m then the the output raster file resolution can be 1, 2 or 5 m but not 3 m.

The same module spdinterp will permit to generate DTMs, DSMs and CHMs by choosing the output model option and the type of point elevation. To interpolate topographic point elevations generates a DTM in the next way:

$ spdinterp --dsm --topo --in NATURAL_NEIGHBOR -f GTiff -b 1 -i QueenElisabeth_example.spd -o DSM.tif

The graphical interface of this tool is located in Processing Toolbox ‣ [Processing] LiDAR ‣ Create digital models:

The interpolation module can produce a DSM

The result can be visualised in QGIS:

Visualisation of a DSM in QGIS

Moreover, importing the spd files generated so far would not show any different result from that imported at the beginning of this tutorial:

$ spdinterp --dtm --topo --in NATURAL_NEIGHBOR -f GTiff -b 1 -i QueenElisabeth_example_mccgrd.spd -o DTM.tif

$ spdinterp --chm --height --in NATURAL_NEIGHBOR -f GTiff -b 1 -i QueenElisabeth_example_mccgrd.spd -o CHM.tif

The interpolation module can produce a DTM

Visualisation of a DTM in QGIS

The interpolation module can produce a CHM

Visualisation of a CHM in QGIS

Generate metrics¶

spdmetrics calculates metrics, which can be simple statistical moments, percentiles of the point height or return amplitude, or even count ratios. Mathematical operators can be applied to either other operators or metric primitives to allow a range of LiDAR metrics to be derived.

Multiple metrics can be calculated at the same time if listed within an XML. This XML file has to be defined a priori with a hierarchical list of metrics and operators. Within the metrics tags a list of metrics can be provided by the metric tag. Within each metric the {field attribute is used to name the raster band or vector attribute. Here is an example of an SPD metrics XML file template containing the maximum, the average, the median, the number of pulses, the canopy cover and the percentile 95th.

References

Thermal Processing¶

Thermal Calibration¶

For calibration of the thermal image we have the following data:

• Thermal image of the study area in Huelva, each pixel of the image represents the temperature value in degrees kelvin. thld_training\huelva\Z1_holmOak\Temperature\t_huelvaZ1.tif
• Set of calibration values stored in a vector file. Each item has the id and the value of collected temperature (in degrees Kelvin). thld_training\huelva\Z1_holmOak\Temperature\t_aois\t_calibration.shp

We proceed to load both the image and the shape to QGIS interface. To load the image in the main menu (Layer - Add Raster Layer ...) to load the shape of points (Layer - Add Vector Layer ...).

Thermal image of Huelva and attribute table containing the field thermal data

We proceed to perform the calibration of the thermal image through Thermolidar tool toolbox [Processing] Thermal - Thermal Calibration.

Thermal calibration tool

We introduce the values as shown in the interface:

Interface of the Thermal Calibration module

We obtain as output calibration file of the thermal image, as shown in the following figure:

Thermal calibration output

Tc - Ta¶

We proceed to subtract the air temperature to the calibrated raster temperatures. We have available the following information.

• Raster temperature calibrated in the previous section. thld_training\huelva\Z1_holmOak\out\thermal_processing\t_calibration.tif
• Air temperature acquired in flight time. In our case we will use the value of temperature 298.15 ºK

We proceed to perform the calibration of the thermal image through Thermolidar tool toolbox [Processing] Thermal - Tc-Ta.

Tc Ta module is located in the “Thermal” submenu of the THERMOLIDAR plugin toolbox

We introduce the input parameters in the user interface:

Interface of the Tc-Ta module

Obtaining as output file the following result:

Output thermal image of Huelva

Data analysis¶

Health condition levels¶

We have the physiology data of Leaf Area Index (LAI) for 25 of the 216 trees inventoried in the area of Huelva, stored in the vector file: thld_training\huelva\Z1_holmOak\FieldData\LAI\lai_huelva.shp.

First, load the shape of polygons in QGIS (Layer - Add vector layer ...)

LAI measurements in Huelva

Before proceeding with the classification of items by level of damage according to several variables taken in the field, we verify that the set of physiological variables follow a normal distribution. For this we use the Shapiro test, located in the toolbox [Analysis] Health Condition Level > Shapiro Test.

Warning

The first time you use these tools, you will need to start QGIS in administrator mode (R install required dependencies)

Shapiro Test¶

Shapire Test is located in the “Health Condition Level” submenu of the THERMOLIDAR plugin toolbox

• Input vector: Vector file that contains information on physiological data.
• Var: Vector’s field to analyze if it follows the normal distribution. In this case the parameter LAI

We introduce the parameters as shown in the following figure:

Interface of the “Shapiro Test” module

Obtaining the following output:

In the example, the p-value is much higher than 0.05, so we conclude that LAI data follow a normal distribution. In the case where p-value is less than 0.05 the data would be discarded, or these should be normalized.

In the case that the variable does not follow a normal distribution, it is necessary to standardize using the [Analysis] Health Condition Level> Standarize

Clustering¶

This tool allows us to group one or more physiological variables according to their degree of similarity between individuals in the sample. In this case we will group by the variable LAI, we have previously verified that follows a normal distribution. This tool creates many groups tool damage level depending on the specified physiological variable. In this case, we will select 3 levels as a function of LAI variable. We can find the tool in [Analysis] Health Condition Level > Clustering

Clustering is located in the “Health Condition Level” submenu of the THERMOLIDAR plugin toolbox

We introduce the values as shown in the following figure:

The vector output file will be stored in the folder thld_training\huelva\Z1_holmOak\out\hcl.shp that will be used subsequently. Each of the trees will be classified as regions of interest to guide the supervised classification of the thermal image.

We obtain results in the dendrogram with the cluster of data:

LAI data grouped into 3 categories health condition, with the following results:

Finally, we must ensure that the groups are significantly different, according to the variables used. This is done through the tool [Analysis] Heath condition level> ANOVA.

ANOVA¶

ANOVA is located in the “Health Condition Level” submenu of the THERMOLIDAR plugin toolbox

Selected as the dependent variable the group that owns each parcel; and as the dependent variable that we want to check if it is significant in the group.

We get the following output:

If the critical level associated with the F statistics (ie, the probability of obtaining values as obtained or older), is less than 0.05, we reject the hypothesis of equal means and conclude that not all the population means being compared are equal. Otherwise, we cannot reject the hypothesis of equality and we cannot claim that the groups being compared differ in their population averages.

Structurally homogenous forest units¶

The analytical purpose of this tool is the definition of structurally homogeneous stands that allow us to minimize the effects of structure on the thermal information, and therefore allow us to obtain related health outcomes woodland.

We start from the metric of the objects to be classified. This vector file was generated at the point of obtaining metric lidar [Processing] LiDAR > Calculate metrics.

• thld_training\huelva\Z1_holmOak\out\metrics.shp

The user must enter the equation by which you want to group the stands, for example according to the equation of the dominant height. In our case, we use the equation obtained from the 95th percentile.

The tool used in this section is located in [Analysis] Structurally homogeneus forest units

SHFU is located in the “Structuraly Homogeneous Forest Units” submenu of the THERMOLIDAR plugin toolbox

In this example, the polygons will be classified into 3 different groups of homogeneity (Cluster parameter).

The Cutoff parameter determines the accuracy of the fit. A value of 0 is of higher precision and larger number of iterations will be needed to reach the final solution.

Interface of the “SHFU” module

We get as output:

Now, we must assign the SHFU group to the objects (regions of interest) described on the Forest Health Level section.

Thus we will use vector files: * thld_training\huelva\Z1_holmOak\out\hcl.shp * thld_training\huelva\Z1_holmOak\out\shfu.shp

We will use the tool QGIS Geoalgorithms > Vector general tools > Join attributes table

We insert the parameters as shown in the following figure:

We will obtain a new output file (thld_training\huelva\Z1_holmOak\out\aois.shp) will use to guide the classification of the thermal image, are of interest to the following fields: * SHFU. Homogeneous group that the object owns. * HCL. Health condition level groups.

Unsupervised pixel-based classification¶

Through this tool raster classify the temperature as many classes as you specify. Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

Interface of the “Unsupervised pixel-based classification” module

We get as output:

Unsupervised object-based classification¶

In the same way that the pixel-oriented, this tool classification raster classify the temperature as many classes as you specify. Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

The difference from the previous tool, is that instead of being classified pixels are classified objects. The classification is made based on the mean value of each of the objects.

Interface of the “Unsupervised object-based classification” module

We get as output:

Supervised pixel-based classification¶

Through this tool the temperature raster will be classified, based on the condition levels defined in the regions of interest (AOIs). * thld_training\huelva\Z1_holmOak\out\aois.shp

Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

Interface of the “Supervised pixel-based classification” module

We get as output:

Supervised object-based classification¶

Through this tool the temperature raster will be classified, based on the condition levels defined in the regions of interest (AOIs). * thld_training\huelva\Z1_holmOak\out\aois.shp

Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

The difference from the previous tool, is that instead of being classified pixels are classified objects. The classification is made based on the mean value of each of the objects.

Interface of the “Supervised object-based classification” module

We get as output:

Thermal Processing¶

Thermal Calibration¶

First, we have a thermal image of the study area. Each digital image value represents the temperature in degrees kelvin.

Thermal image of Sierra de los Filabres

Several values of temperature field of invariant surfaces (black and white cloth) have been collected, and GPS position of each sample.

Attribute table of shapefile containing the field thermal data.

We introduce the input parameters in the user interface:

Interface of the Thermal Calibration module

As a result the software generates a calibrated image of the study area.

Tc - Ta¶

We introduce the input parameters in the user interface:

• Input layer: Calibrated thermal raster (generated in the previous section)
• Air temperature: Constant air temperature measure at flight time. In this case, the air temperature is considered to 293.15 degrees kelvin

Interface of the Tc-Ta module

Output thermal image of Sierra de los Filabres

Data analysis¶

Health condition levels¶

In the following example we have acquired several variables physiology field (LAI), for a number of control plots in Filabres.

LAI measurements in Filabres

Before proceeding with the classification of items by level of damage according to several variables taken in the field, we verify that the set of physiological variables follow a normal distribution. For this we use the Shapiro test, located in the toolbox [Analysis] Health Condition Level > Shapiro Test.

Warning

The first time you use these tools, you will need to start QGIS in administrator mode (R install required dependencies)

Shapiro Test¶

• Input vector: Vector file that contains information on physiological data.
• Var: Vector’s field to analyze if it follows the normal distribution. In this case the parameter lai_LICOR2

Interface of the “Shapiro Test” module

Obtaining the following output:

In the example, the p-value is much higher than 0.05, so we conclude that LAI data follow a normal distribution. In the case where p-value is less than 0.05 the data would be discarded, or these should be normalized.

In the case that the variable does not follow a normal distribution, it is necessary to standardize using the [Analysis] Health Condition Level> Standarize

Standarize¶

We verified that the lai_LICOR2 variable follows a normal distribution.

• Input vector: Vector file that contains information on physiological data.
• Var: Shapefile’s field to standardize. In this case the parameter lai_LICOR2

Clustering¶

This tool allows us to group one or more physiological variables according to their degree of similarity between individuals in the sample. In this case we will group by the variable lai_LICOR2, we have previously verified that follows a normal distribution. This tool creates many groups tool damage level depending on the specified physiological variable. In this case, we will select 3 levels as a function of LAI variable.

LAI data grouped into 3 categories health condition, with the following results:

Finally, we must ensure that the groups are significantly different, according to the variables used. This is done through the tool [Analysis] Heath condition level> ANOVA.

ANOVA¶

Selected as the dependent variable the group that owns each parcel; and as the dependent variable that we want to check if it is significant in the group.

We get the following output:

If the critical level associated with the F statistics (ie, the probability of obtaining values as obtained or older), is less than 0.05, we reject the hypothesis of equal means and conclude that not all the population means being compared are equal. Otherwise, we cannot reject the hypothesis of equality and we cannot claim that the groups being compared differ in their population averages.

Structurally homogenous forest units¶

The analytical purpose of this tool is the definition of structurally homogeneous stands that allow us to minimize the effects of structure on the thermal information, and therefore allow us to obtain related health outcomes woodland.

The user must enter the equation by which you want to group the stands, for example according to the equation of the dominant height. In our case, we use the equation obtained from the 95th percentile. In this example, the polygons will be classified into 3 different groups of homogeneity.

Interface of the “SHFU” module

We get as output:

Forest Health Classification¶

Unsupervised pixel-based classification¶

Through this tool raster classify the temperature as many classes as you specify. Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

Interface of the “Unsupervised pixel-based classification” module

We get as output:

Unsupervised object-based classification¶

In the same way that the pixel-oriented, this tool classification raster classify the temperature as many classes as you specify. Temperature classification is performed based on homogeneous units, so many output raster have been defined as SHFU be obtained.

In our case, we obtain three raster classified. Each raster has associated many categories defined temperature.

The difference from the previous tool, is that instead of being classified pixels are classified objects. The classification is made based on the mean value of each of the objects.

Interface of the “Unsupervised object-based classification” module

We get as output:

USAGE¶

spdclearclass  -o <String> -i <String>  [-c <unsigned int>] [-r <unsigned  int>] [--] [--version] [-h]

NAME¶

Clear the classification of an SPD file: spdclearclass

USAGE¶

spdcopy  -o <String> -i <String> [-c  <unsigned int>] [-r <unsigned int>]  [--] [--version] [-h]

NAME¶

Makes a copy of an indexed SPD file: spdcopy

USAGE¶

spddecomp  -o <String> -i <String> [-a]  [-d <float>] [-w <uint_fast32_t>]  [-e <uint_fast32_t>] [-n] [-t  <uint_fast32_t>] [-c <unsigned  int>] [-r <unsigned int>] [--]  [--version] [-h]

NAME¶

Decompose full waveform data to create discrete points: spddecomp

USAGE¶

spddefheight  {--interp|--image} -o  <String> [-e <String>] -i <String>  [--idxres <float>] [--thinres  <float>] [--ptsperbin  <uint_fast16_t>] [--thin]  [--tpsnopts <uint_fast16_t>]  [--tpsRadius <float>]  [--stdDevRadius <float>]  [--largeRadius <float>]  [--smallRadius <float>]  [--stddevThreshold <float>] [--in  <TIN_PLANE|NEAREST_NEIGHBOR|NATURAL_NEIGHBOR|STDEV_MULTISCALE|TPS_RAD|TPS_PTNO>] [-b <float>]  [--overlap <uint_fast16_t>] [-c  <unsigned int>] [-r <unsigned int>]  [--] [--version] [-h]

NAME¶

Define the height field within pulses and points: spddefheight

USAGE¶

spddefrgb  {--define|--stretch} -o  <String> [--image <String>] -i  <String> [--independ] [--coef  <float>] [--blue <uint_fast16_t>]  [--green <uint_fast16_t>] [--red  <uint_fast16_t>] [-c <unsigned  int>] [-r <unsigned int>]  [--stddev] [--linear] [--]  [--version] [-h]

NAME¶

Define the RGB values on the SPDFile: spddefrgb

USAGE¶

spddeftiles  {-t|-e} [-i <String>] [-o  <String>] [--ymax <double>] [--xmax  <double>] [--ymin <double>] [--xmin  <double>] [--overlap <double>]  [--ysize <double>] [--xsize  <double>] [--] [--version] [-h]

NAME¶

Tools for defining a set of tiles: spddeftiles

USAGE¶

spdelevation  {--constant <double>|--variable <string>} {--add|--minus} -o <String> -i <String>  [-c <unsigned int>] [-r <unsigned  int>] [--] [--version] [-h]

NAME¶

Alter the elevation of the pulses: spdelevation

USAGE¶

spdextract  -o <String> -i <String>  [--max] [--min] [--return <ALL|FIRST|LAST|NOTFIRST|FIRSTLAST>]  [--class <unsigned int>] [-b  <float>] [-c <unsigned int>] [-r  <unsigned int>] [--] [--version]  [-h]

NAME¶

Extract returns and pulses which meet a set of criteria: spdextract

USAGE¶

spdgrdtidy  {--negheights} -o <String> -i  <String> [-c <unsigned int>] [-r  <unsigned int>] [--] [--version]  [-h]

NAME¶

Attempt to tidy up the ground return classification: spdgrdtidy

USAGE¶

spdinfo  [--] [--version] [-h] <string>  ...

NAME¶

Print header info for an SPD File: spdinfo

USAGE¶

spdinterp  {--dtm|--chm|--dsm|--amp}  {--topo|--height|--other} -o  <String> -i <String> [--rbflayers  <unsigned int>] [--rbfradius  <double>] [--idxres <float>]  [--thinres <float>] [--ptsperbin  <uint_fast16_t>] [--thin]  [--tpsnopts <uint_fast16_t>]  [--tpsRadius <float>]  [--stdDevRadius <float>]  [--largeRadius <float>]  [--smallRadius <float>]  [--stddevThreshold <float>] [--in  <TIN_PLANE|NEAREST_NEIGHBOR|NATURAL_NEIGHBOR|STDEV_MULTISCALE|TPS_RAD|TPS_PTNO|RBF>] [-f  <string>] [-b <float>] [--overlap  <uint_fast16_t>] [-c <unsigned  int>] [-r <unsigned int>] [--]  [--version] [-h]

NAME¶

Interpolate a raster elevation surface: spdinterp

USAGE¶

spdlastest  {-p|-c|-f} -i <String> [-n  <unsigned int>] [-s <unsigned int>]  [--] [--version] [-h]

NAME¶

Print data pulses from a LAS file - for debugging: spdlastest

USAGE¶

spdmaskgen  -o <String> -i <String> [-f  <string>] [-b <float>] [-c  <unsigned int>] [-r <unsigned int>]  [-p <unsigned int>] [--]  [--version] [-h]

NAME¶

Generate a binary mask for the an input SPD File: spdmaskgen

USAGE¶

spdmccgrd  -o <String> -i <String>  [--thresofchangemultireturn]  [--class <uint_fast16_t>]  [--median] [--thresofchange  <float>] [--filtersize  <uint_fast16_t>] [--interpnumpts  <uint_fast16_t>] [--interpmaxradius  <float>] [--stepcurvetol <float>]  [--mincurvetol <float>]  [--initcurvetol <float>]  [--scalegaps <float>]  [--numofscalesbelow  <uint_fast16_t>]  [--numofscalesabove  <uint_fast16_t>] [--initscale  <float>] [--overlap  <uint_fast16_t>] [-b <float>] [-c  <unsigned int>] [-r <unsigned int>]  [--] [--version] [-h]

NAME¶

Classifies the ground returns using the multiscale curvature algorithm: spdmccgrd

USAGE¶

spdmerge  -o <String> [--keepextent] [-s  <std::string>] [--classes  <uint_fast16_t>] ...  [--returnIDs  <uint_fast16_t>] ...  [--source]  [--ignorechecks] [-c] [-r  <std::string>] [-p <std::string>]  [--wavebitres <8BIT|16BIT|32BIT>]  [-x <FIRST_RETURN|LAST_RETURN|START_WAVEFORM|END_WAVEFORM|ORIGIN|MAX_INTENSITY|UNCHANGED>] -f <SPD|ASCIIPULSEROW|ASCII|FWF_DAT|DECOMPOSED_DAT|LAS|LASNP|LASSTRICT|DECOMPOSED_COO|ASCIIMULTILINE>  [--] [--version] [-h] <std::string>  ...

NAME¶

Merge compatable files into a single non-indexed SPD file: spdmerge

USAGE¶

spdmetrics  {--image|--vector|--ascii}  [-v <String>] -m <String> -o  <String> -i <String> [-f <string>]  [-b <float>] [-c <unsigned int>]  [-r <unsigned int>] [--]  [--version] [-h]

NAME¶

Calculate metrics : spdmetrics

USAGE¶

spdoverlap  {-c|-s} [-o <String>] [--]  [--version] [-h] <string> ...

NAME¶

Calculate the overlap between UPD and SPD files: spdoverlap

USAGE¶

spdpffgrd  -o <String> -i <String>  [--gdal <string>] [--class  <uint_fast16_t>] [--morphmin]  [--image] [-m <uint_fast16_t>] [-f  <uint_fast32_t>] [--tophatscales  <bool>] [-t <uint_fast32_t>] [-s  <uint_fast32_t>] [-k  <uint_fast32_t>] [--grd <float>]  [--overlap <uint_fast16_t>] [-b  <float>] [-c <unsigned int>] [-r  <unsigned int>] [--] [--version]  [-h]

NAME¶

Classifies the ground returns using a parameter-free filtering algorithm: spdpffgrd

USAGE¶

spdpmfgrd  -o <String> -i <String>  [--gdal <string>] [--class  <uint_fast16_t>] [--image]  [--medianfilter <uint_fast16_t>]  [--nomedian] [--grd <float>]  [--maxelev <float>] [--initelev  <float>] [--slope <float>]  [--maxfilter <uint_fast16_t>]  [--initfilter <uint_fast16_t>]  [--overlap <uint_fast16_t>] [-b  <float>] [-c <unsigned int>] [-r  <unsigned int>] [--] [--version]  [-h]

NAME¶

Classifies the ground returns using the progressive morphology algorithm: spdpmfgrd

USAGE¶

spdpolygrd  {--global|--local} -o  <String> -i <String> [--class  <uint_fast16_t>] [--iters <int>]  [--degree <int>] [--grdthres  <float>] [-b <float>] [--overlap  <uint_fast16_t>] [-c <unsigned  int>] [-r <unsigned int>] [--]  [--version] [-h]

NAME¶

Classify ground returns using a surface fitting algorithm: spdpolygrd

USAGE¶

spdprofile  -o <String> -i <String> [-f  <std::string>] [-b <float>] [-c  <unsigned int>] [-r <unsigned int>]  [-n <unsigned int>] [-t <unsigned  int>] [-m <float>] [-w <unsigned  int>] [--order <unsigned int>]  [--smooth] [--] [--version] [-h]

NAME¶

Generate vertical profiles: spdprofile

USAGE¶

spdproj  {--proj4 <string>|--proj4pretty  <string>|--image <string>|--imagepretty <string>|--spd  <string>|--spdpretty <string>|--epsg <int>|--epsgpretty <int>|--shp <string>|--shppretty  <string>} [--] [--version] [-h]

NAME¶

Print and convert projection strings: spdproj

USAGE¶

spdrmnoise  -o <String> -i <String>  [--grellow <float>] [--grelup  <float>] [--rellow <float>]  [--relup <float>] [--abslow  <float>] [--absup <float>] [-c  <unsigned int>] [-r <unsigned int>]  [--] [--version] [-h]

NAME¶

Remove vertical noise from LiDAR datasets: spdrmnoise

USAGE¶

spdstats  {--image|--overall} -o <String>  -i <String> [-f <string>] [-b  <float>] [-c <unsigned int>] [-r  <unsigned int>] [--] [--version]  [-h]

NAME¶

Provides statistics the point and pulse density of an SPD file: spdstats

USAGE¶

spdsubset  -o <String> -i <String> [--num  <uint_fast32_t>] [--start  <uint_fast32_t>] [--shpfile  <string>] [--ignorez]  [--ignorerange] [--txtfile  <string>] [--spherical] [--height]  [--ranmax <double>] [--ranmin  <double>] [--zenmax <double>]  [--zenmin <double>] [--azmax  <double>] [--azmin <double>]  [--hmax <double>] [--hmin <double>]  [--zmax <double>] [--zmin <double>]  [--ymax <double>] [--ymin <double>]  [--xmax <double>] [--xmin <double>]  [--] [--version] [-h]

NAME¶

Subset point cloud data: spdsubset

USAGE¶

spdthin  -o <String> -i <String> [-n  <unsigned int>] [-c <unsigned int>]  [-r <unsigned int>] [--]  [--version] [-h]

NAME¶

Thin a point cloud to a defined bin spacing: spdthin

USAGE¶

spdtileimg  {-m|-c} [-i <String>] [-t  <String>] [-o <String>] [-w  <String>] [-b <double>] [-r  <double>] [-f <String>] [--]  [--version] [-h]

NAME¶

Tools for mosaicing raster results following tiling: spdtileimg

USAGE¶

spdtiling  {--all|--extract|--extractcore|--tilespdfile|--builddirs|--rmdirs|--upxml|--xml2shp} [--deleteshp]  [-u] [-d] [--useprefix]  [--usedirstruct] [-c <Unsigned  int>] [-r <Unsigned int>] [--wkt  <String>] [-i <String>] [-o  <String>] -t <String> [--]  [--version] [-h]

NAME¶

Tools for tiling a set of SPD files using predefined tile areas: spdtiling

USAGE¶

spdtranslate  -o <String> -i <String>  [--keepextent] [--pulseversion  <unsigned int>] [--pointversion  <unsigned int>] [--wavenoise  <float>] [--Oz <float>] [--Oy  <double>] [--Ox <double>]  [--defineOrigin] [--tly <double>]  [--tlx <double>] [--defineTL]  [--keeptemp] [--convert_proj]  [--output_proj <string>]  [--input_proj <string>] [-b  <float>] [-c <unsigned int>] [-r  <unsigned int>] [-s <string>] [-t  <string>] [--scan] [--polar]  [--spherical] [--wavebitres <8BIT|16BIT|32BIT>] [-x <FIRST_RETURN|LAST_RETURN|START_WAVEFORM|END_WAVEFORM|ORIGIN|MAX_INTENSITY|UNCHANGED>] --of <SPD|UPD|ASCII|LAS|LAZ> --if <SPD|ASCIIPULSEROW|ASCII|FWF_DAT|DECOMPOSED_DAT|LAS|LASNP|LASSTRICT|DECOMPOSED_COO|ASCIIMULTILINE> [--] [--version]  [-h]

NAME¶

Convert between file formats: spdtranslate

USAGE¶

spdversion  [--] [--version] [-h]  <string> ...

NAME¶

Prints version information: spdversion

USAGE¶

spdwarp  {--shift|--warp} -o <String> -i  <String> [-c <unsigned int>] [-r  <unsigned int>] [-y <float>] [-x  <float>] [-g <std::string>]  [--order <unsigned int>] [-p <>]  [-t <POLYNOMIAL|NEAREST_NEIGHBOR|TRIANGULATION>] [--] [--version]  [-h]

NAME¶

Interpolate a raster elevation surface: spdwarp