Install the requirements: numpy, opencv-python, opencv-contrib-python Convert the input video to .avi format

Run the main.py script:

• Imports the algorithm's API using "from FastHorizonAlg import FastHorizon"
• Instantiates a class instance
• Invoke the horiozn detection method on the .avi video

Docker

Build:

docker build -t fast_detector -f docker/Dockerfile .

Run:

docker run -it --rm --name vsnt_opencv  -e DISPLAY=$DISPLAY  -e QT_X11_NO_MITSHM=1 -v /tmp/.X11-unix:/tmp/.X11-unix -v ${HOME}/Documents/git/A_fast_horizon_detector_and_a_new_annotated_dataset_for_maritime_video_processing:/fast_detector -v ${HOME}/Documents/datasets/TMD:/datasets/TMD fast_detector

docker run -it --rm --name vsnt_opencv  -e DISPLAY=$DISPLAY  -e QT_X11_NO_MITSHM=1 -v /tmp/.X11-unix:/tmp/.X11-unix -v ${HOME}/git/A_fast_horizon_detector_and_a_new_annotated_dataset_for_maritime_video_processing:/fast_detector -v ${HOME}/Documents/datasets/TMD:/datasets/TMD fast_detector

Convert TMD .npy files to .csv

python3 tmd_npy_to_csv.py