An image processing utility built upon OpenCV to generate motion heat maps from video-based sources. Originally used to analyse foot traffic in a large open area.

Note that this application is purely inspired by intel-iot-devkit/motion-heatmap-cpp.

Also, the updated version of this script uses two old repos, as inspiration: (they dont work tho)

• Heat-Mapper: https://github.com/gugarosa/heat_mapper
• Video Parser: https://github.com/Vibhootikishor/Converting-Frames-to-Video

Some optimization, some basic improvements and stuff to overall better suit this to my purpose. Saved it as a repo just in case I needed it in the future.

Install all the pre-needed requirements using:

pip install -r requirements.txt

This repository is composed of four scripts that assist one in producing a heat map of a motion-based video.

To run the scripts, first copy and paste a video onto the project folder. Then, run the following command:

python heat_mapping.py <video_name>

Add the --write flag if you want the frames to be written to disk. Video rendering requires the frames to be written to disk.

The first step is to test whether the video is displaying correctly. Use the following script and pass a source file that contains a video as its argument:

python video_displaying.py -h

Note that -h invokes the script helper, which assists users in employing the appropriate parameters.

Afteward, it is possible to construct a heat map based on the video's motion. Essentially, the algorithm removes the static background and sequentially accumulates the motion throughout the frames. At the end of the process, the resulting images will composed the motion-based masks and the individual frames.

python heat_mapping.py -h

The --write flag defines whether the video will be displayed as an inline screen or saved to disk.

Finally, with every frame outputted as an image, it is possible to reconstruct them back to a video. Use the following script to achieve such a purpose:

In the old repository, the video rendering script was sort of broken. So, I decided to make a new one, using Vibhootikishor's implementation.

This is done automatically. Remember, the script only terminates when the video is rendered successfully.

Note that the video is configured to render as an .mp4. If necessary, change the codec according to the desired extension.

If you're using Windows, good job.

If Mac uhh sorry good luck tho.

There is literally no support for this repo I am currently getting screwed over by uni I don't have time for your bullshit.