Traffic monitoring with yolov7

This repository contains the implementation of YOLOv7-based vehicle detection and counting system using Python. The system is designed to detect and count vehicles in real-time from video or image inputs, providing valuable traffic data for road analysis. It also supports the use of a CSI camera on Jetson Nano.

Features

• Real-time vehicle detection and counting
• Support for video and image inputs
• Efficient and accurate vehicle detection using YOLOv7
• Easy-to-use Python API for integrating into your projects
• CSI camera support for Jetson Nano

Test samples

for more samples check https://www.youtube.com/channel/UCqv1qrgRdgH8aCm-YG47qPA

Commands

1. python detect.py --weights 982-11mb.pt --view-img --source video.mp4 --show-fps

2. python detectpy --weights 982-47mb.pt --view-img --nosave --source 0 --show-fps --conf 0.5 --seed 2 --track

3. python detect.py --weights 982-11mb.pt --view-img --nosave --source image.jpg --show-fps --conf 0.5

İmportant Note

To run it on jetson nano with CSI camera (considering all configurations have done) utils/dataset.py should be changed to frame is taken by a CSI camera. it is recommended to check face_detect.py from https://github.com/JetsonHacksNano/CSI-Camera/blob/master/face_detect.py

Referances

• https://github.com/WongKinYiu/yolov7
• https://github.com/RizwanMunawar/yolov7-object-tracking/
• https://github.com/krisnarengga/yolov7-vehicle-counting