Giter Site home page Giter Site logo

mxnet-cnn-plate-recognition's Introduction

基于CNN的OCR车牌识别



所需环境

  • Python2.7
  • Mxnet
  • Numpy
  • Opencv

操作步骤

1.生成车牌sample[^code]

python genPlate.py 100 /Users/shelter/plate_100

参数1:生成车牌的数量
参数2:生成车牌存放的地址

2.训练CNN模型[^code]

python train.py 

3.预测车牌准确率

#随机生成100张车牌图片
python genPlate.py 100 /Users/shelter/test

#批量预测测试图片准确率
python test.py /Users/shelter/test

##输出结果示例
output:
预测车牌号码为:津 K 4 2 R M Y
输入图片数量:100
输入图片行准确率:0.72
输入图片列准确率:0.86

参考资料

1.http://blog.csdn.net/relocy/article/details/52174198
2.https://github.com/szad670401/learning-dl/tree/master/mxnet/ocr


新增功能@11.8

1.增加生成黄色车牌功能 genYellowPlate.py
 - 黄色车牌包括大车、农用车、教练车、摩托车、试验车,目前只能生成正常车牌(教练车、一行、7位数),摩托车、大车暂时还未开发,明天开发@11.9
2.增加生成绿色车牌功能 gen GreenPlate.py
 - 绿色车牌主要是新能源车牌,背景图片与蓝色车牌有区别,位数为8位,与传统的7位蓝色车牌不同。当然,绿底的车牌也包括几种:小型新能源车号牌(非纯电动、纯电动)、大型新能源车号牌(非纯电动、纯电动),前者底色为全绿,上部会有少许白色,后者全绿色,左边两个字符的底色为黄色。目前只开发了前者的车牌生成,后者还未找到合适的背景图。


待开发功能

1.最近做了个yolov3的车牌检测的项目,上周跑了个demo,这周针对性优化开发了三天,准确率为99.25%,还有不少的优化空间。过段时间把检测+识别一体化,端到端输出
2.白底车牌和黑底车牌的检测和识别也有待增加。目前数据集太少,也得生成一些比较真实的数据集,可以尝试用GAN的方法生成更加逼真的车牌

mxnet-cnn-plate-recognition's People

Contributors

huxiaoman7 avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

mxnet-cnn-plate-recognition's Issues

我需要这个cnn-ocr-model-symbol.json文件,才能通过吗?

raise MXNetError(py_str(_LIB.MXGetLastError())) mxnet.base.MXNetError: [16:38:56] C:\projects\mxnet-distro-win\mxnet-build\3rdpa rty\dmlc-core\src\io\local_filesys.cc:166: Check failed: allow_null LocalFileSy stem::Open "/data/mxnet/cnn-ocr-model-symbol.json": No such file or directory

can you help me sovle this problem?

Traceback (most recent call last):
File "E:/LicensePlate_Recognition/mxnet-cnn-plate-recognition/train.py", line 304, in
train();
File "E:/LicensePlate_Recognition/mxnet-cnn-plate-recognition/train.py", line 284, in train
model = mx.mod.FeedForward(ctx=devs, #使用GPU来跑

  • AttributeError: module 'mxnet.module' has no attribute 'FeedForward'

关于yolo定位车牌

我用了yolo检测车牌,有一个问题,就是用yolo弄数据集只能是正方形或者长方形,那么很多图片车牌都是不规则的,训练完,框都画出来但是画出来的框会有些比较大,有些比较小一点,你是怎么处理这个问题

两行车牌

你好这个可以做有两行字符的车牌识别嘛,比如香港的车牌

黄牌数据咨询

作者,你好,正常的黄牌数据应该是字是黑色的,而作者代码生成的黄牌是白字,请问如何修改才能让字变成黑色

why u add so many ; in the code? that not c....

关于新能源车牌识别问题

请问作者,关于新能源绿色车牌的识别,训练代码要如何考虑设计?是否能够将7位车牌和8位车牌放在一起训练?

车牌生成

cv2.error: OpenCV(4.1.0) /io/opencv/modules/imgproc/src/resize.cpp:3718: error: (-215:Assertion failed) !ssize.empty() in function 'resize'

我用的python3 使用车牌生成报这个错误,是因为opencv对2和3区别所导致的吗?还是哪里的问题呢?请问

用gpu运行

请问这个项目要怎么用GPU运行,我把devs改成mx.gpu()会报错
mxnet.base.MXNetError: [16:44:09] src/imperative/imperative.cc:78: Operator _zeros is not implemented for GPU.

Accuraccy=0.0000?

小姐姐您好!
为什么我在训练你的网络时得到到正确率全为0呢?
2020-07-04 11:28:13,605 Epoch[0] Batch [0-50] Speed: 178.25 samples/sec Accuracy=0.000000
2020-07-04 11:28:15,878 Epoch[0] Batch [50-100] Speed: 176.06 samples/sec Accuracy=0.000000
2020-07-04 11:28:18,151 Epoch[0] Batch [100-150] Speed: 175.98 samples/sec Accuracy=0.000000
2020-07-04 11:28:20,427 Epoch[0] Batch [150-200] Speed: 175.67 samples/sec Accuracy=0.000000
2020-07-04 11:28:22,717 Epoch[0] Batch [200-250] Speed: 174.67 samples/sec Accuracy=0.000000
2020-07-04 11:28:24,989 Epoch[0] Batch [250-300] Speed: 176.13 samples/sec Accuracy=0.000000
2020-07-04 11:28:27,266 Epoch[0] Batch [300-350] Speed: 175.59 samples/sec Accuracy=0.000000
2020-07-04 11:28:29,552 Epoch[0] Batch [350-400] Speed: 175.05 samples/sec Accuracy=0.000000
2020-07-04 11:28:31,825 Epoch[0] Batch [400-450] Speed: 175.98 samples/sec Accuracy=0.000000
2020-07-04 11:28:34,115 Epoch[0] Batch [450-500] Speed: 174.60 samples/sec Accuracy=0.000000

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.