README.md

How to use segmentmytif

Segment (georeferenced) raster data in an interactive fashion. Retrain models in seconds. Only small amounts of labeled data necessary because of our use of pretrained base models as feature extractors.

The project setup is documented in project_setup.md.

Installation

To install segmentmytif from GitHub repository, do:

git clone [email protected]:DroneML/segmentmytif.git
cd segmentmytif
python -m pip install .

Logging

The application writes logs to the 'logs' dir, which will be created if it doesn't exist yet. Messages printed to the screen (stdout) are stored in info.log for later reference. More detailed information, such as input data shapes and value distributions, are written to debug.log.

Train a feature extraction model

To train a feature extraction model run the script "train_model.py" in this repo:

python ./src/segmentmytif/utils/train_model.py -r ../monochrome_flair_1_toy_dataset_flat/ --train_set_limit 10

This assumes a 'flat', grayscale, version of the FLAIR1 dataset is present at the selected root location.

root
- train
    - input
        - IMG_061946_0.tif
        - IMG_061946_1.tif
        - ...
    - labels
        - MSK_061946_0.tif
        - ...    

Use the script 'monochromize.py' to create greyscale (single band) tifs for every multiband tif in a source folder:

python ./src/segmentmytif/utils/monochromize.py -i ../flair_1_toy_dataset/ -o ../monochrome_flair_1_toy_dataset/

Credits

This package was created with Copier and the NLeSC/python-template.