Video Processing

Note

All options below, will assume you have pulled the docker image, and tagged it as tracking-image to simplify the following commands (so they're not platform dependant). Please see the docker quickstart section for more details.

To run only the Radar processing section of the tracking program, you can disable video processing. This can be beneficial if you only want to collect radar data, or do not have a video source available.

This will walk you through how to modify the container running tracking, but also applied to the local developer environment as well.

Running ONLY Video Processing

This configuration is useful if you'd only like to collect video data, and are not worried about tracking anything. This is useful for testing video data collection and ensure the model is running. It's also useful to see the Yolo performance by itself

Run the Container Interactively

LinuxJetson5

docker run -it tracking-image

docker run --ipc=host --runtime=nvidia -it

Run the tracking program, with radar processing and all tracking disabled.
```
python3 tracking.py --skip-radar --skip-tracking
```

Running with Video and Tracking Processing

This configuration is useful if you'd only like to collect video data and see how it influences the tracking.

Run the Container Interactively

LinuxJetson5

docker run -it tracking-image

docker run --ipc=host --runtime=nvidia -it tracking-image

Run the tracking program, with video processing and radar tracking disabled.
```
python3 tracking.py --skip-radar
```

Disable Saving Video Data

When running the tracking program, include the --disable-save-video option.

python3 tracking.py --disable-save-video