Wetland classification using country-wide ALS data

This repository aims to collect all the scripts which were used for wetland classification based on Airborne Laser Scanning data within the e-EcoLiDAR project.

The content of the directories

• analysis:

These scripts contain the processing steps related to analyze, convert and classify the ALS data.

• figures:

Visualization and camera ready figure plot for the geobia conference paper.

• grassgis_process:

GRASS GIS batch scripts for LiDAR data segmentation.

• laserchicken_process:

Using laserchicken point cloud analysis software for extracting features related to ALS data.

• workflow:

Bash file for executing the scripts from this repository and build a consistent workflow for processing the ALS data.

• testdata:

Small example dataset for testing purpose.

Installation requirements

Generally, the workflow was tested under Windows 10 operation system. The workflow can be run under Linux operation system, however, the GRASS GIS section should be adapted.

1. Install Git: https://gitforwindows.org/ Follow the instructions from here: https://escience-academy.github.io/2017-09-06-git-github/#setup

2. Install Anaconda (most preferably the latest version with Python 3): https://conda.io/docs/user-guide/install/index.html

3. Install GRASS GIS (version 7.2 or above): https://grasswiki.osgeo.org/wiki/Installation_Guide (OSGeo4W installer can be accessed during QGIS installation: https://qgis.org/en/site/forusers/download.html)

Install laserchicken software

1. Get the software (GitHub repository: https://github.com/eEcoLiDAR/laserchicken)

a.) Clone the repository (official release)

git clone https://github.com/eEcoLiDAR/laserchicken.git

b.) Download as zip file without using git

2. Install

Within the command line (using the right python version) go to this directory and run the following:

pip install .

If shapely failed than install it separately:

conda install -c conda-forge shapely

Install addons for GRASS GIS

Two addons is required to install for running this workflow. For the installation after starting up GRASS GIS from command line (grass74 -gtext) you can install addons with the following commands:

g.extension extension=r.neighborhoodmatrix
g.extension extension=i.segment.uspo

Usage

Before running the entire workflow you should check that in the bash file ( wetland_classification_from_lidar/workflow/Workflow_geobia_confpaper.sh ) the defined absolute paths are correct. The script can be executed from Git bash (MINGW64 environment) command line using Windows 10 operation system.

• work_folder: location of test files on the local computer (wetland_classification_from_lidar/testdata directory, the results will be saved in this directory too)
• script_path: location of the wetland_classification_from_lidar on the local computer
• path_of_laserchicken: location of laserchicken repository on the local computer
• grass_path: location of the installed GRASS GIS bin directory on the local computer
• grass_mapset: before running the workflow user should set up a mapset with Descartes coordinate system (run from GRASS GIS command line separately grass74 -gtext and then follow the instructions of the software and do not specify specific projection)

The valid_polygon is provided within this repository not the original one (just an example course digitalization from my side). It can be only used for testing. Also the test data is a small tile of the study area from AHN2 laser campaign only for demonstration (we do not own this dataset).

bash <filepath>/Workflow_geobia_confpaper.sh