Project in IKT213-G21H Machine Vision to automatically detect cracks in roads.
Run trainer.py
to train a model on images in /tools/out
and /tools/test
. Images not included.
View modle_trainer.ipynb
to get a more detailed view of how we trained our models.
Run tester.py
to preform a quick test of the model on a few example images in /Examples/
.
Run analyzeVideo.py
to analyze a video file and save every detected crack into /Examples/out/
.
Run analyzeCamera.py
to analyze video stream from a the Jetson Nano camera and save every detected crack into /Examples/out/
.
To create image sets from videos quickly use videoToImages.py
in /tools/
. It takes two parameters <input-video> <output-folder>
.