# **Automatic Number Plate Detection system**
github:[https://github.com/201918010332Thomas/ANPR](https://github.com/201918010332Thomas/ANPR)
**Software environment requirements: python >=3.6 pytorch >=1.7**
## **GUI program:**
```
python anpr.py
```
## License plate detection training
1. **Dataset**
This project uses open source datasets CCPD and CRPD.
The dataset label format is YOLO format:
```
label x y w h pt1x pt1y pt2x pt2y pt3x pt3y pt4x pt4y
```
The key points are in order (top left, top right, bottom right, bottom left).
The coordinates are all normalized, where x and y are the center point divided by the width and height of the image, w and h are the width and height of the box divided by the width and height of the image, ptx and pty are the key point coordinates divided by the width and height.
2. **Modify the data/widerface.yaml file**
```
train: /your/train/path #This is the training dataset, modify to your path.
val: /your/val/path #This is the evaluation dataset, modify to your path.
# number of classes
nc: 2 #Here we use 2 categories, 0 single layer license plate 1 double layer license plate.
# class names
names: [ 'single','double']
```
3. **Train**
```
python train.py --data data/widerface.yaml --cfg models/yolov5n-0.5.yaml --weights weights/plate_detect.pt --epoch 250
```
The result exists in the run folder.
4. **Detection model onnx export**
To export the detection model to onnx, onnx sim needs to be installed. **[onnx-simplifier](https://github.com/daquexian/onnx-simplifier)**
```
1. python export.py --weights ./weights/plate_detect.pt --img 640 --batch 1
2. onnxsim weights/plate_detect.onnx weights/plate_detect.onnx
```
**Using trained models for detection**
```
python detect_demo.py --detect_model weights/plate_detect.pt
```
## License plate recognition training
The training link for license plate recognition is as follows:
[License plate recognition training](https://github.com/201918010332Thomas/CRNN_LPR)
#### **The results of license plate recognition are as follows:**
![Image](image/README/test_12.jpg)
## Arrange
1.**onnx demo**
The onnx model can be found in [onnx model](https://pan.baidu.com/s/1UmWN2kpRP96h2cM6Pi-now), with extraction code: ixyr
python onnx_infer.py --detect_model weights/plate_detect.onnx --rec_model weights/plate_rec.onnx --image_path imgs --output result_onnx
2.**tensorrt**
Deployment can be found in [tensorrt_plate](https://github.com/we0091234/chinese_plate_tensorrt)
3.**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/deepcam-cn/yolov5-face)
* [https://github.com/meijieru/crnn.pytorch](https://github.com/meijieru/crnn.pytorch)
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
软件环境要求:python >=3.6 pytorch >=1.7 GUI程序: python anpr.py 车牌检测培训 数据 该项目使用开源数据集 CCPD 和 CRPD。 数据集标注格式为 YOLO 格式: label x y w h pt1x pt1y pt2x pt2y pt3x pt3y pt4x pt4y 关键点按顺序排列(左上、右上、右下、左下)。 坐标都是归一化的,其中 x 和 y 是中心点除以图像的宽度和高度,w 和 h 是盒子的宽度和高度除以图像的宽度和高度,ptx 和 pty 是关键点坐标除以宽度和高度。 修改 data/widerface.yaml 文件 train: /your/train/path #This is the training dataset, modify to your path. val: /your/val/path #This is the evaluation dataset, modify to your path. # number of classes nc: 2
资源推荐
资源详情
资源评论
收起资源包目录
基于YOLOv5人脸和CRNN模型的自动车牌识别系统 (1309个子文件)
events.out.tfevents.1678598819.My_HW.13752.0 5KB
events.out.tfevents.1679650292.My_HW.19860.0 5KB
events.out.tfevents.1679629790.My_HW.18384.0 5KB
events.out.tfevents.1678601065.My_HW.16808.0 5KB
events.out.tfevents.1678599805.My_HW.18008.0 5KB
events.out.tfevents.1678598536.My_HW.17216.0 5KB
events.out.tfevents.1679629350.My_HW.19704.0 5KB
events.out.tfevents.1678600022.My_HW.10976.0 5KB
events.out.tfevents.1678625856.My_HW.3680.0 5KB
events.out.tfevents.1678596274.My_HW.3368.0 5KB
events.out.tfevents.1679629672.My_HW.3348.0 5KB
events.out.tfevents.1679628229.My_HW.21552.0 5KB
events.out.tfevents.1679971433.My_HW.19560.0 5KB
events.out.tfevents.1679629231.My_HW.18304.0 5KB
events.out.tfevents.1679627519.My_HW.22424.0 5KB
events.out.tfevents.1678552366.My_HW.7736.0 5KB
events.out.tfevents.1678598445.My_HW.3064.0 5KB
events.out.tfevents.1678625235.My_HW.9712.0 5KB
events.out.tfevents.1679628440.My_HW.21660.0 5KB
events.out.tfevents.1678552531.My_HW.12960.0 5KB
events.out.tfevents.1678625437.My_HW.7680.0 5KB
events.out.tfevents.1679649715.My_HW.23476.0 5KB
events.out.tfevents.1679972384.My_HW.16648.0 5KB
events.out.tfevents.1679629588.My_HW.22348.0 5KB
events.out.tfevents.1679629066.My_HW.22176.0 5KB
events.out.tfevents.1678625387.My_HW.17072.0 5KB
events.out.tfevents.1679627940.My_HW.22368.0 5KB
events.out.tfevents.1678597932.My_HW.3260.0 5KB
events.out.tfevents.1678598757.My_HW.12652.0 5KB
events.out.tfevents.1679629463.My_HW.20344.0 5KB
events.out.tfevents.1679625269.My_HW.18684.0 5KB
events.out.tfevents.1678601378.My_HW.7804.0 5KB
events.out.tfevents.1678599045.My_HW.11392.0 5KB
events.out.tfevents.1678599325.My_HW.16924.0 5KB
events.out.tfevents.1678600756.My_HW.15576.0 5KB
events.out.tfevents.1678436242.My_HW.10828.0 5KB
events.out.tfevents.1678437852.My_HW.16996.0 5KB
events.out.tfevents.1678374414.My_HW.18892.0 5KB
events.out.tfevents.1678432061.My_HW.14048.0 5KB
events.out.tfevents.1678436078.My_HW.8604.0 5KB
events.out.tfevents.1678374252.My_HW.4984.0 5KB
events.out.tfevents.1678431896.My_HW.4232.0 5KB
events.out.tfevents.1678432265.My_HW.18440.0 5KB
events.out.tfevents.1678377349.My_HW.4664.0 5KB
events.out.tfevents.1678373970.My_HW.8624.0 5KB
events.out.tfevents.1678435762.My_HW.17640.0 5KB
events.out.tfevents.1678323734.My_HW.20340.0 5KB
events.out.tfevents.1678323653.My_HW.19360.0 5KB
events.out.tfevents.1678377980.My_HW.17780.0 5KB
events.out.tfevents.1678437569.My_HW.7536.0 5KB
events.out.tfevents.1678435403.My_HW.6724.0 5KB
events.out.tfevents.1678431019.My_HW.12196.0 5KB
events.out.tfevents.1678374262.My_HW.15032.0 5KB
events.out.tfevents.1678436129.My_HW.13160.0 5KB
events.out.tfevents.1678435915.My_HW.11076.0 5KB
events.out.tfevents.1678437446.My_HW.10096.0 5KB
events.out.tfevents.1678432551.My_HW.15992.0 5KB
events.out.tfevents.1678437768.My_HW.7964.0 5KB
events.out.tfevents.1678429728.My_HW.14368.0 5KB
events.out.tfevents.1678370143.My_HW.18048.0 5KB
events.out.tfevents.1678378576.My_HW.15716.0 5KB
events.out.tfevents.1678378622.My_HW.14448.0 5KB
events.out.tfevents.1678365104.My_HW.4112.0 5KB
events.out.tfevents.1679975500.My_HW.22868.0 173B
events.out.tfevents.1679975721.My_HW.720.0 173B
events.out.tfevents.1679975551.My_HW.14084.0 173B
events.out.tfevents.1679974183.My_HW.20816.0 173B
events.out.tfevents.1679975468.My_HW.2216.0 173B
events.out.tfevents.1679625243.My_HW.21176.0 88B
events.out.tfevents.1678551938.My_HW.14840.0 88B
events.out.tfevents.1678323549.My_HW.8972.0 40B
box_overlaps.c 269KB
Dockerfile 821B
bbox.cp38-win_amd64.exp 745B
single_blue.jpg 1.81MB
test_1.jpg 1.26MB
xue.jpg 999KB
single_green.jpg 903KB
minghang.jpg 584KB
hongkang1.jpg 571KB
test_12.jpg 549KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
labels_correlogram.jpg 411KB
共 1309 条
- 1
- 2
- 3
- 4
- 5
- 6
- 14
资源评论
hakesashou
- 粉丝: 6537
- 资源: 1658
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功