add image dataset description

[luiskuhn]

• Provide technical description of dataset, e.g. image channels, spatial size, acquisition parameters, etc. (in READMEs and zenodo datasets)
• Document which channel is used as input for segmentation (in README and activity diagram of https://github.com/qbic-pipelines/root-tissue-segmentation-core/blob/development/README.rst)
• keep in mind: Can we train the model using other channels, or combinations?

[luiskuhn]

This is fixed in: https://github.com/qbic-pipelines/root-tissue-segmentation-core/pull/5