Tutorials

Active Fire Detection

The following tutorial will demonstrate how to run an existing docker Image that seamlessly detects fire events across the United States using GOES satellite data.

  1. Go to a deployed GeoNEX website, for example: http://13.57.231.201:8080

  2. Enter XXXX for both Username and password and click login

    • The dashboard page should become visible, this is where you control running Images and access their output among other things.

  3. On the left menu, click on the GeoNEX Containers tab

  4. Under the Action column, click Action then Run for the row with the following Image:
    • geonex/active_fire_go16abi15
    • This will lead you to a new page for selecting the Container name (Running Image name) configurations.
  5. Make sure the following is entered:
    • Runner Name: My Fire Detection
    • Instance type: m5d.4xlarge
    • Running type: real-time runner
    • Data Source: GOES16
    • Docker Command Line: docker run --rm geonex/active_fire_go16abi15:1.5 -s 20183112017182 -r RadC -u s3://geonex-yunfeng/GO16_ABI15_DEMO
  6. Click Run Docker
    • This will launch the Image as a docker container in a new AWS instance and will take you to the Real-time runner page (which is also accessible from the left-menu), where you can see the status of all Images previously run.
    • If all goes well during that process the status column should eventually show running
    • Every time the real-time container you just run executes, it creates a timestamped task entry.
  7. To view the tasks executed by the running container My Fire Detection: under the Action tab, click Action, then Tasks.
    • All tasks run in the same instance designated by the instance-id column.
  8. Once the state column of the tasks generated show finished, you can select __Action__ then Outputs to see the outputs. Each task should take less than 5 minutes to complete.
  9. This Image generates 4 output files:
    • .json file that contains: Fire events coordinates and their properties (temperature..) in geojson format
    • .log file: Detailing a timestamped task log
    • .nc file: Raw output of fire detection script (fire mask map in netcdf format)
    • .png file: A visual output showing a map with overlaid fire data.
  10. As long as the container is running, these tasks will be executed as soon as GOES maps are available.
    • WARNING: Make sure to stop the container once you no longer need real-time outputs to save on costs.
  11. A sample display of the output is shown below: img