GuidosToolbox Workbench

This document provides usage instructions for the image analysis module GWB (GuidosToolbox Workbench), here implemented as a Jupyter dashboard on SEPAL.

Introduction

In 2008, the GuidosToolbox (GTB, Vogt and Riitters 2017) was developed as a graphical user interface (GUI) to Morphological Spatial Pattern Analysis (MSPA) of raster data (Soille and Vogt 2009). The GTB has since been enhanced with numerous modules for analysis of landscape objects, patterns, and networks, and specialized modules for assessing fragmentation and restoration. The GuidosToolbox Workbench (GWB) provides the most popular GTB modules as a set of command-line applications for 64bit Linux systems. In the following, we describe the implementation of GWB on the SEPAL platform as a Jupyter dashboard based on the GWB CLI tool.

Presentation

To launch the app please follow the SEPAL registration steps and then move to the SEPAL Apps dashboard (purple wrench icon on the left side panel), search for and click on GWB ANALYSIS.

SEPAL dashboard

The application should launch itself in the About section, allowing to select the tool you want to use.

Note

If this is the first time you use the app, you will actually see the following popup:

licence

This licence needs to be accepted to use the GWB modules. It is also available in the section Licence of the app. If you don’t want to accept this Licence, just close the app tab.

General structure

The application is strucured as followed:

On the left side you will find a navigation drawer that you can open and close using the Hamburger button.

Tip

On small devices such as tablet or phones, the navigation drawer will be hidden by default. Click the Hamburger button (highlighted by the red square) to show the full extent of the app.

small screen without drawer small screen with drawer

Each name in the list corresponds to one GWB module, presented individually in the next sections. By clicking on it you will display the panels relative to the function you want to use.

presentation of the structure

Danger

All GWB modules require categorical raster input maps in data type unsigned byte (8bit), with discrete integer values within [0, 255] byte. Any other data format will raise an error.

Modules

Each module is presented individually. You can directly jump to the module of interest by clicking on the related link under the section Modules in the right panel of this page and this documentation will guide you through the respective processing steps.

ACC

This module will conduct the Accounting analysis. Accounting will label and calculate the area of all foreground objects. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Accounting product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

  • special background 1 (optional)

  • special background 2 (optional)

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 4 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background. If you specify special background, it will be treated separately in the analysis (e.g. water, buildings).

Select the parameters

You will need to select parameters for your computation:

acc params

Note

These parameters can be used to perform the default computation:

  • Foreground connectivity: 8

  • spatial pixel resolution: 25

  • area thresholds: 200 2000 20000 100000 200000

  • options: default

Foreground connectivity

This set the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image
Spatial pixel resolution

Set the spatial pixel resolution of your image in meters. It is only used for the summary.

Area thresholds

Set up to 5 area thresholds (measured in pixels).

Options

Two computation options are available:

  • stats + image of viewport (default)

  • stats + images of ID, area, viewport (detailed)

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/acc/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_acc.tif

  • <raster_name>_bin_map_acc.csv

  • <raster_name>_bin_map_acc.txt

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_acc.tif

DIST

This module will conduct the Euclidean Distance analysis. Each pixel will show the shortest distance to the foreground boundary. Pixels inside a foreground object have a positive distance value while background pixels have a negative distance value. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Distance product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 2 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background.

Select the parameters

You will need to select parameters for your computation:

dist params

Note

These parameters can be used to perform the default computation:

  • Foreground connectivity: 8

  • Options: Euclidian Distance only

Foreground connectivity

This set the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image
Options

Two computation options are available:

  • compute the Euclidian Distance only

  • compute the Euclidian Distance and the Hysometric Curve

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/dist/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_dist.tif

  • <raster_name>_bin_map_dist.txt

  • <raster_name>_bin_map_dist_hist.png

  • <raster_name>_bin_map_dist_viewport.tif

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_dist_hmc.png https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_dist.tif

FAD

This module will conduct the fragmentation analysis at five fixed observation scales. Because forest fragmentation is scale-dependent, fragmentation is reported at five observation scales, which allows different observers to make their own choice about scales and threshold of concern. The change of fragmentation across different observation scales provides additional interesting information. Fragmentation is measured by determining the Forest Area Density (FAD) within a shifting, local neighborhood. It can be measured at pixel or patch level. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Fragmentation product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

  • special background 1 (optional)

  • special background 2 (optional)

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 4 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background. If you specify special background, it will be treated separately in the analysis (e.g. water, buildings).

Warning

The special background 2 is the non-fragmenting background (optional), see the Fragmentation product sheet for details.

Select the parameters

You will need to select parameters for your computation:

acc params

Note

These parameters can be used to perform the default computation:

  • Foreground connectivity: 8

  • Computation precision: float-precision

  • Options: per-pixel density, color-coded into 6 fragmentation classes (FAD)

Foreground connectivity

This sets the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image
Computation precision

Set the precision used to compute your image. Float precision (default) will give more accurate results compared to rounded byte but will also take more computing resources and disk space.

Options

Three computation options are available:

  • FAD: per-pixel density, color-coded into 6 fragmentation classes

  • FAD-APP2: average per-patch density, color-coded into 2 classes

  • FAD-APP5: average per-patch density, color-coded into 5 classes

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/fad/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_fad_<class_number>.tif

  • <raster_name>_bin_map_fad_barplot.png

  • <raster_name>_bin_map_fad_mscale.csv

  • <raster_name>_bin_map_fad_mscale.tif

  • <raster_name>_bin_map_fad_mscale.txt

  • <raster_name>_bin_map_fad_mscale.sav

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_fad_barplot.png https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_fad_mscale.tif

FRAG

This module will conduct the fragmentation analysis at a user-selected observation scale. This module and its option are similar to fad but allow the user to specify a single (or multiple) specific observation scale. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Fragmentation product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

  • special background 1 (optional)

  • special background 2 (optional)

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 4 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background. If you specify special background, it will be treated separately in the analysis (e.g. water, buildings).

Warning

The special background 2 is the non-fragmenting background (optional), see the Fragmentation product sheet for details.

Select the parameters

You will need to select parameters for your computation:

acc params

Note

These parameters can be used to perform the default computation:

  • Foreground connectivity: 8

  • Spatial pixel resolution: 25

  • Computation precision: float-precision

  • Windows size: 23

  • Options: average per-patch density, color-coded into 2 classes (FAD-APP2)

Foreground connectivity

This sets the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image
Spatial pixel resolution

Set the spatial pixel resolution of your image in meters. Only use for the summary.

Window size

Set up to 10 observation windows sizes (in pixels).

Options

Three computation options are available:

  • FAD: per-pixel density, color-coded into 6 fragmentation classes

  • FAD-APP2: average per-patch density, color-coded into 2 classes

  • FAD-APP5: average per-patch density, color-coded into 5 classes

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/frag/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_frag_fad-<option>_<class>.tif

  • <raster_name>_bin_map_frag.csv

  • <raster_name>_bin_map_frag.txt

  • <raster_name>_bin_map_frag.tif

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the FAD-APP2 option on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_fad-app2_23.tif

LM

This module will conduct the Landscape Mosaic analysis at a user-selected observation scale. The Landscape Mosaic measures land cover heterogeneity, or human influence, in a tri-polar classification of a location accounting for the relative contributions of the three land cover types Agriculture, Natural, Developed in the area surrounding that location. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Landscape Mosaic product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/clc3class.tif file (3 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • dominant land cover 1 (Agriculture)

  • dominant land cover 2 (Natural)

  • dominant land cover 3 (Developed)

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 3 classes

Select the parameters

You will need to select parameters for your computation:

lm params

Note

This parameter can be used to perform the default computation:

  • window size: 23

Window size

Set the square window size (in pixels). It should be an odd number in [3, 5, …501]. with \(kdim\) being the window size, which is related to the observation scale by the following formula:

\[obs_scale = (pixres * kdim)^2 / 10000\]

with

  • \(obs_scale\) in hectare

  • \(pixres\) in meters

  • \(kdim\) in pixels

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/lm/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_lm_23.tif

  • <raster_name>_bin_map_lm_23_103class.tif

  • <raster_name>_bin_map_heatmap.csv

  • <raster_name>_bin_map_heatmap.png

  • <raster_name>_bin_map_heatmap.sav

  • heatmap_legend.png

  • lm103class_legend.png

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the clc3classes.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/lm103class_legend.png https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/clc3class_lm_23.tif

MSPA

Warning

If your are considering using the MSPA tool, keep in mind that the process is relatively complex and provide a lot of informmation (up to 25 classes). If you are only interested in fragmentation and/or less than 6 classes, please consider using FRAG or SPA.

This module will conduct the Morphological Spatial Pattern Analysis. MSPA analyses shape and connectivity and conducts a segmentation of foreground patches in up to 25 feature classes. The result are spatially explicit maps and tabular summary statistics. Details on the methodology and input/output options can be found in the Morphology product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 2 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background.

Select the parameters

You will need to select parameters for your computation:

mspa params

Note

These parameters can be used to perform the default computation:

  • Foreground connectivity: 8 (default)

  • Edge width: 1

  • Transition: True

  • Intext: True

Foreground connectivity

This sets the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image
Edge width

Define the width (measured in pixels) of the Core-boundaries (Edges and Perforations).

Transition

Select if you want to show transition pixels, where connecting pathways go through edges/perforations (transition=1 (true), default) or not (transition=0).

Intext

Select if you want to distinguish MSPA classes and Holes laying within Core objects (intext=1 (true), default) or not (intext=0).

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/mspa/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_<4 params>.tif

  • <raster_name>_bin_map_<4 params>.txt

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/mspalegend.gif https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_8_1_1_1.tif

P223

This module will conduct the Density (P2), Contagion (P22) or Adjacency (P23) analysis of foreground (FG) objects at a user-selected observation scale (Riitters et al. (2000)). The result are spatially explicit maps and tabular summary statistics. The classification is determined by measurements of forest amount (P2) and connectivity (P22) within the neighborhood that is centered on a subject forest pixel. P2 is the probability that a pixel in the neighborhood is forest, and P22 is the probability that a pixel next to a forest pixel is also forest.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

  • special background (for P23 only)

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 3 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background. If you specify special background, it will be treated separately in the analysis (e.g. water, buildings)

Select the parameters

You will need to select parameters for your computation:

p223 params

Note

These parameters can be used to perform the default computation:

  • Window size: 27

  • Computation precision: Float (default)

  • Algorithm: FG-Density

Window size

Set the square window size (in pixels) It should be an odd number in [3, 5, …501] with \(kdim\) being related to the observation scale by the following formula:

\[obs_scale = (pixres * kdim)^2 / 10000\]

with

  • \(obs_scale\) in hectare

  • \(pixres\) in meters

  • \(kdim\) in pixels

Computation precision

Set the precision used to compute your image. Float precision (default) will give more accurate results compared to rounded byte but will also take more computing resources and disk space.

Algorithm

The P223 module can run: FG-Density (P2), FG-Contagion (P22), or FG-Adjacency (P23)

P223 will provide a color-coded image showing [0,100]% for either FG-Density, FG-Contagion, or FG-Adjacency masked for the Foreground cover. Use the alternative options to obtain the original spatcon output without normalisation, masking, or color-coding.

Tip

For original spatcon output ONLY: Missing values are coded as 0 (rounded byte), or -0.01 (float precision). For all output types, missing indicates the input window contained only missing pixels.

Tip

For FG-Contagion and FG-Adjacency output ONLY, missing also indicates the input window contained no foreground pixels (there was no information about foreground edge).

For all output types, \(rounded byte = (float precision * 254) + 1\)

You’ll find the options displayed with the following names in the dropdown menu:

  • FG-Density (FG-masked and normalised)

  • FG-Contagion (FG-masked and normalised)

  • FG-Adjacency (FG-masked and normalised)

  • FG-Density (original spatcon output)

  • FG-Contagion (original spatcon output)

  • FG-Adjacency (original spatcon output)

  • FG-Shannon (original spatcon output)

  • FG-SumD (original spatcon output)

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/p223/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_p<option>_<window>.tif

  • <raster_name>_bin_map_p<option>_<window>.txt

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the P2 (Foreground-Density) option on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_p2_27.tif

PARC

This module will conduct the parcellation analysis. This module provides a statistical summary file (txt/csv- format) with details for each unique class found in the image as well as the full image content: class value, total number of objects, total area, degree of parcellation. Details on the methodology and input/output options can be found in the Parcellation product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/clc3classes.tif file (3 classes).

The first step requires to select your image in your SEPAL folder. The image must be a categorical tif raster.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

upload 0 classes

Select the parameters

You will need to select parameters for your computation:

parc params

Note

This parameter can be used to perform the default computation:

  • Foreground connectivity: 8

Foreground connectivity

This set the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/parc/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_parc.csv

  • <raster_name>_bin_map_parc.txt

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the clc3classes.tif file.

Class

Value

Count

Area[pixels]

APS

AWAPS

AWAPS/data

DIVISION

PARC[%]

1

1

45

2.44893e+06

54420.7

2.07660e+06

1.27136e+06

0.152039

1.19374

2

2

164

5840.73

82557.6

19770.0

0.913812

17.7426

3

3

212

2798.07

19008.4

0.783919

11.0897

8-connected Parcels:

421

4000000

9501.19

1310139.4

0.672465

8.07904

RSS

This module will conduct the Restoration Status Summary analysis. It will calculate key attributes of the current network status, composed of foreground (forest) patches and it provides the normalized degree of network coherence. The result are tabular summary statistics. Details on the methodology and input/output options can be found in the Restoration Planner product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 2 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background.

Select the parameters

You will need to select parameters for your computation:

rss params

Note

This parameters can be used to perform the default computation:

  • Foreground connectivity: 8

Foreground connectivity

This set the foreground connectivity of your analysis:

  • 8 neighbors (default) will use every pixel in the vicinity (including diagonals)

  • 4 neighbors only use the vertical and horizontal one

connectivity image

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/rss/, for example:

  • <raster_name>_bin_map.tif

  • rss<connectivity>.txt

  • rss<connectivity>.csv

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using the default parameters on the example.tif file.

FNAME

AREA

RAC[%]

NR_OBJ

LARG_OBJ

APS

CNOA

ECA

COH[%]

example_bin_map.tif

428490.00

42.860572

2850

214811

150.34737

311712

221292.76

51.644789

SPA

This module will conduct the Simplified Pattern Analysis. SPA analyses shape and conducts a segmentation of foreground patches into 2, 3, 5, or 6 feature classes. The result are spatially explicit maps and tabular summary statistics. GWB_SPA is a simpler version of GWB_MSPA. Details on the methodology and input/output options can be found in the Morphology product sheet.

Setup the input image

Tip

You can use the default dataset to test the module. Click on the Download test dataset button on the top of the second panel. By clicking on this button, the following two files will be added to your downloads folder:

  • example.tif: 0 byte - Missing, 1 byte - Background, 2 byte - Foreground

  • clc3class.tif: 1 byte - Agriculture, 2 byte - Natural, 3 byte - Developed

download sample dataset

Once the files are downloaded, follow the normal process using the downloads/example.tif file (2 classes).

The first step requires to reclassify your image. Using the reclassifying panel, select your image in your SEPAL folder.

Warning

If the image is not in your SEPAL folders but in your local computer consider reading the exchange file with SEPAL page of this documentation.

The dropdown menus will list the discrete values of your raster input image. Select each class in your image and place them in one of the following categories:

  • background

  • foreground

Every class that is not set to a reclassifying category will be considered as “missing data” (0 byte).

upload 2 classes

Tip

For forest analysis, set forest as foreground and all the other classes as background.

Select the parameters

You will need to select parameters for your computation:

spa params

Note

This parameter can be used to perform the default computation:

  • number of pattern classes: 2: Small & linear features (SLF), Coherent

Number of pattern classes

Set the number of pattern classes you want to compute:

  • 2: Small & linear features (SLF), Coherent

  • 3: Core, Core-Openings, Margin

  • 5: Core, Core-Openings, Edge, Perforation, Margin

  • 6: Core, Core-Openings, Edge, Perforation, Islet, Margin

Run the analysis

Once your parameters are all set you can launch the analysis. The blue rectangle will display information about the computation. Upon completion, it will turn to green and display the computation log.

information logs

The resulting files are stored in the folder module_results/gwb/spa/, for example:

  • <raster_name>_bin_map.tif

  • <raster_name>_bin_map_spa<number of classes>.tif

  • <raster_name>_bin_map_spa<number of classes>.txt

Danger

If the rectangle turns red, carefully read the information in the log. For example, your current instance may be too small to handle the file you want to analyse. In this case, close the app, open a bigger instance and run your analysis again.

Here is the result of the computation using SPA2 (2 classes) on the example.tif file.

https://raw.githubusercontent.com/openforis/sepal-doc/master/docs/source/img/cli/gwb/example_spa2.tif