ecognition/linux_es repository overview

⁠Trimble eCognition GRID Engine Server for Linux

Trimble eCognition software is used by GIS professionals, remote sensing experts & data scientists to automate geospatial data analytics. Users can design feature extraction solutions to transform geo-data into geo-information.

Trimble eCognition GRID consists of these components:

• Engine Server
• Job Scheduler (ecognition/linux_js⁠ or ecognition/win_js⁠)
• Command-line tools. For job submitting, monitoring and etc. (ecognition/linux_cle⁠ or ecognition/win_cle⁠)

You should only use eCognition GRID if your rule set is using one of the following process algorithms:

• submit scenes for analysis
• read subscene statistics
• create scene copy
• create scene subset
• create scene tiles
• delete scenes In all other cases, eCognition Command-line Engine should be enough.(ecognition/linux_cle⁠ or ecognition/win_cle⁠)

The Linux version of Docker image for Trimble eCognition Engine Server. You will need a valid Trimble eCognition Server license to run it.

⁠Prerequisite

This component is processing jobs from job queue hosted by JobScheduler component. You will need a running instance of JobScheduler component.

To use the eCognition Engine Server requires a valid eCognition Server license and the latest eCognition License Server⁠.

Note: If License Server is running on a different machine than container, you will need to make sure license server ports (default: 27000-27009) and vendor daemon (random by default) are open. You can set those ports via license server admin console using your favorite browser: http://[LicenseServer IP address]:8090 (or http://localhost:8090⁠ from within License Server machine). in Administration tab you can set those in "Vendor Daemon Configuration" and "Server Configuration" sections.

⁠Run

To use the container you can run it and it will automatically connect to JobScheduler. For this you need to specify JobScheduler and LicenseServer IP addresses:

docker pull ecognition/linux_es:10.5.0
docker run --rm -d --name ecog_linux_es -e JOB_SCHEDULER=myJobSchedulerIP:8184 -e LM_LICENSE_FILE=@myLicenseIP --network=grid_network -m 8g --cpus=2 ecognition/linux_es:10.5.0

The above example assumes network grid_network being available. You can create it with command like this:

docker network create -d nat --subnet=172.28.0.0/16  --gateway=172.28.5.254 grid_network

Optionally, you may want to mount shared folder for data sharing and explicit log file path and detail level, by adding these parameters to docker run command:

-v /myShare:/mnt -e ECOG_CONFIG_logging="log path=/mnt/logs;trace level=Detailed"

Note: It is recommended to give at least 4GB of RAM for container as it might be resource intensive (docker -m 4GB .. will do it)

⁠Configuration via environment

While running container any configuration setting (found in config/eCognition.cfg) can be set via environment variable use the following notation: ECOG_CONFIG_{config-group}="key1=value1;key2=value2;.."

for example, to set log file location and trace level (this overrides appropriate <logging> values in eCognition.cfg):

docker run -it -e ECOG_CONFIG_logging="log path=/mnt/logs;trace level=Detailed" ...

⁠Limitations

Linux version of the Engine Server does not support point cloud processing. For point cloud based workflow, please use windows version of the server.