rapideye-cloud-detection

The purpose of this repository is to isolate the cloud detection process for rapideye scenes. In order to do so, we use a in-house approach that consist on detecting anomalies by measuring the expected behavior of the values in a particular band and the ones that are actually obtained.

Disclosure

This procedure is merely experimental, it is a proposal, and by no means it is a final product. It makes no sense to take it as THE MADMex cloud masking algorithm.

How to Create the Cloud Mask

First of all we need to create the docker image:

docker build -t rapideye-clouds .

Once we have the docker image built, we start the docker container attaching the current directory, and the directory containing the rapideye scene:

docker run -it \
       -v <path to rapideye directory>:<path to rapideye directory inside docker> \
       -v $(pwd):/data \
       rapideye-clouds \
       python main.py <path to rapideye directory inside docker>

Windows

In Windows, Docker is only allowed to mount files that are found in the directory:

C:\\Users

So our working directory must be in that directory, for the sake of example, I created a directory called example:

C:\\Users\example

Recalling that the paths for Windows must be prepend by a "/" we have:

docker run -it \
       -v /<path to rapideye directory>:<path to rapideye directory inside docker> \
       -v /$(pwd):/data \
       rapideye-clouds \
       python main.py <path to rapideye directory inside docker>

For instance, if our working directory is in "C:\Users\example\l3a":

docker run -it \
       -v /c/Users/example/l3a/:/rapideye/ \
       -v /$(pwd):/data \
       rapideye-clouds \
       python main.py /rapideye/

Remark

It can be the case that the RAM is not enough for this process, if that is the case, settings in VirtualBox must be changed to something around 8192Mb.