post-discharge-channel-segmentation

Model

post-discharge-channel.pt is a YOLO26 segmentation model trained to identify post-discharge channels in shadowgraph images. The predicted masks can be used to measure channel area, dimensions, deformation, displacement, and lifetime in high-speed experimental recordings.

Inference Script

post_discharge_channel_segmentation_yolo26.py processes PNG and JPG frames in alphabetical order. It saves frames with segmentation masks, creates an MP4 video, calculates mask area versus time, exports the measurements to CSV, generates a plot with a moving-median trend, and creates a 4×4 grid of selected frames.

Place the model, script, and input folder together:

post_discharge_channel_segmentation_yolo26.py
post-discharge-channel.pt
frames/

Install the dependencies and run the script:

pip install -U ultralytics opencv-python numpy matplotlib scipy
python post_discharge_channel_segmentation_yolo26.py

All configurable parameters are located at the beginning of the script.

Citation

If you use this model, dataset, or accompanying software in your research, please cite one or more of the following publications, selecting those most relevant to your work:

• https://doi.org/10.1016/j.actaastro.2025.09.091
• https://doi.org/10.26089/NumMet.v24r217
• https://doi.org/10.1615/JFlowVisImageProc.2025058234
• https://doi.org/10.1016/j.actaastro.2023.11.021
• https://doi.org/10.26583/sv.15.4.01