License Plate Recognition

Tech Stack

• Python
• HTML/CSS/Javascript
• SQLite3 Database
• Flask
• YOLOv7
• OpenCV, EasyOCR, Pytorch, Numpy

To run my script:

Clone the yolov7 repository. Place the detect-plates.py script and best.pt weights file in the yolov7 directory. Then run this command (all the commands below involve usage of CUDA, Nvidia GPU, to use CPU refer the documentation in the yolov7 github)

  python detect-plates.py --weights best.pt --conf 0.25 --img-size  640 --source path/to/image --device 0

For internal camera

  python detect-plates.py --weights best.pt --conf 0.25 --img-size  640 --source 0 --device 0

For external camera

  python detect-plates.py --weights best.pt --conf 0.25 --img-size  640 --source 1 --device 0

There are many other customizations in the flags, please go through yolov7's github repository for all of them.

Link to Detection Model

Google Drive Link

Images and Videos of Testing

back.mp4

direction.tracking.mp4

Licnese.Plate.Tracking-1.mp4

Plate.IDs.mp4

YOLOv7

YOLOv7 was made by @WongKinYiu.

YOLOv7's Repository:

• https://github.com/WongKinYiu/yolov7.git