Vehicle-and-Human-Detection

General Approach

Uses a cv2 while ret is True loop to process entire video

• call object detection function (creates bounding boxes)
• call counter function (check if bounding box intersects ROI)
• call drawing functions (ROI line, bounding boxes, counter text)
• return output video

Command Line Interface

In the main directory, run main.py then answer all prompts. python3 main.py You can rename the source video in main.py to your own input.

Input Video Requirements

Input must be 30 FPS for accurate results. Lower resolution videos process signficantly faster.

Supported Python and Tensorflow Versions

The program was implemented running Python 3.8.10 and Tensorflow version 2.4

License Information

Licensed under GNU General Public License v3.0

Please credit Jose Martinez and Wren Priest if you found this repository helpful!