Chinese_license_plate_detection_recognition

What's New

2022.12.04 Check the vehicle and license plate together to see the vehicle system here

yolov8 license plate detection + identification

yolov7 license plate detection + identification

Android NCNN

connect

The model is trained from the public data set, and requires a model with higher accuracy, or please add V for business cooperation.

Can do student graduation and coursework, etc.

wechat: we0091234 (Note the purpose of the visit)

The most complete license plate recognition algorithm supports 12 Chinese license plate types

Environment requirements: python >=3.6 pytorch >=1.7

Image test demo:

Run detect_plate.py directly or run the following command line:

 python detect_plate.py --detect_model weights/plate_detect.pt  --rec_model weights/plate_rec_color.pth --image_path imgs --output result

Test folder imgs, save the result and then enter the result folder

Video test demo 2.MP4 extraction code: 41aq

 python detect_plate.py --detect_model weights/plate_detect.pt  --rec_model weights/plate_rec_color.pth --video 2.mp4

The video file is 2.mp4 and save it as result.mp4

License plate inspection training

The license plate detection training link is as follows:

License plate inspection training

License plate recognition training

The license plate recognition training link is as follows:

License plate recognition training

Supports as follows:

• 1. Single-line blue card
• 2. Single yellow card
• 3. New energy license plate
• 4. White police license plate
• 5. Coach license plate
• 6. Armed Police License Plate
• 7. Double-layer yellow card
• 8. Double-layer white card
• 9. Embassy license plate
• 10. Hong Kong, Macao and Guangdong Z brand
• 11. Double-layer green card
• 12. Civil aviation license plate

deploy

1. Android NCNN

2. onnx demo Baidu network disk: k874

 python onnx_infer.py --detect_model weights/plate_detect.onnx  --rec_model weights/plate_rec_color.onnx  --image_path imgs --output result_onnx

3. See tensorrt_plate for tensorrt deployment

4. openvino demo version 2022.2

 python openvino_infer.py --detect_model weights/plate_detect.onnx --rec_model weights/plate_rec.onnx --image_path imgs --output result_openvino

References

• https://github.com/deepcam-cn/yolov5-face
• https://github.com/Sierkinhane/CRNN_Chinese_Characters_Rec

More

qq group: 769809695 (newly opened in the third group) 871797331 (full) 837982567 (second group is full) Ask