Coder Social home page Coder Social logo

hngynjy / alpr-1 Goto Github PK

View Code? Open in Web Editor NEW

This project forked from youwh-piri/alpr

3.0 1.0 0.0 163.86 MB

基于4种轻量级深度卷积网络的无场景约束全自动车牌识别,轻量级车牌检测,轻量级车牌识别,pyqt5可视化界面

Python 97.31% Shell 2.69%

alpr-1's Introduction

ALPR

基于4种轻量级深度卷积网络的无场景约束全自动车牌识别

车辆检测器控制系统开启状态,提前加载模型,补光灯控制光照,识别控制道闸

车辆检测+车辆分类+车牌定位+车牌修复(强光)+车牌识别

检测:yolov3_tiny(darknet)突出速度

分类:b-cnn(resnet-18)突出准确率

定位:retina-net(mobilenetv1)突出速度

修复:dcgan(TODO)

识别:lpr-net(cnn + ctc)速度与准确率

本次提交以pyqt5界面形式展示 图片名称可省略.jpg

更换车辆检测网络为仅使用yolov3_tiny(pytorch版本)

分类功能暂时不做,修复功能待完善

检测对应文件夹:yolov3_tiny_car_det

size gtx titan x(fps)
33M 220

定位对应文件夹:plate_location

利用RetinaFace进行迁移学习实现的车牌检测、车牌四角定位、车牌矫正对齐

可参考我的博客链接,RETINAFACE论文笔记

https://www.cnblogs.com/ywheunji/p/12285421.html

数据

地址

https://github.com/detectRecog/CCPD

performance

使用mobilenet0.25作为骨干网时,模型大小仅为1.7M

model inference@gtx 1060(ms)
1.7M 0.5-

识别对应文件夹:lprnet_Plate_Recognition

可参考我的博客链接,lprnet论文笔记

https://www.cnblogs.com/ywheunji/p/12268340.html

performance

  • include blue/green license plate.
  • test images number is 27320.
model personal test imgs(%) inference@gtx 1060(ms)
1.7M 96.0+ 0.5-

环境:

  • python3
  • openv-python 3.x
  • pytorch 1.1
  • pyqt5.9
  • imutils
  • Pillow

打开可视化界面:

直接运行camershow.py文件即可进入pyqt的ui界面,支持视频,单张,摄像头模式

目前完成状态:

基本功能建造完成,待优化cpu版本的速度

优化暂无时间做,若追求速度,可直接把车辆检测去掉,基本不影响准确率,可达到低性能设备实时检测效果

参考文献

http://openaccess.thecvf.com/content_ECCV_2018/papers/Sergio_Silva_License_Plate_Detection_ECCV_2018_paper.pdf

https://github.com/lyl8213/Plate_Recognition-LPRnet

https://github.com/gm19900510/Pytorch_Retina_License_Plate

https://github.com/CaptainEven/Vehicle-Car-detection-and-multilabel-classification

https://blog.csdn.net/weixin_43008870/article/details/86496263

http://www.szjiuding.com/237.html

https://github.com/eriklindernoren/PyTorch-YOLOv3

alpr-1's People

Contributors

youwh-piri avatar

Stargazers

 avatar 陈十三 avatar  avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.