Install requirements using:
pip install -r requirements.txt
Run using:
$ python celere.py --weights-b damage_assessment/buildings.pt \
--weights-d damage_assessment/damage_classify.pt \
--source data/images/DamagedBuildings/
Or call from other python script:
celere.run(weights_b='damage_assessment/buildings.pt',
weights_d='damage_assessment/damage_classify.pt',
source='data/images/DamagedBuildings/')
The repository is a fork of ultralytics/yolov5. The principal modifications are:
- Addition of damage_assessment folder, including:
- Pretrained model files
- scoring_logic.py containing helper functions
- train.py and detect.py as evidence of the custom classifier implementation
- ClassDetect.py and classify/ClassClassify.py to enable import of the models as classes (necessary for efficient multiple calls)
- celere.py Main script to access functionality
- annotate_image.py to annotate individual images (handy for the report)
- several minor changes to existing files for data export and to fit requirements