OpenCV实现车牌识别,OCR分割,ANN神经网络

Posted 晴堂

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了OpenCV实现车牌识别,OCR分割,ANN神经网络相关的知识,希望对你有一定的参考价值。

主要步骤:
准备车牌单个字符图像作为神经网络分类器的训练数据,越多越好。当然需要对每幅图像提取特征,这里使用的是水平和垂直累计直方图和缩小后的图像信息。
获取车牌图像,这里的车牌图像已经完成抠图,并且是灰度图像。
将车牌图像中每个字符分割成单一图像(OCR类实现)。
提取分割出的字符图像特征信息,并使用分类识别字符(OCR类实现)。

 程序运行过程:

              

                             原始带有车牌的图片

                         

                           抠图并输入的车牌图片

                             

                      二值化并分割成单个字符图片

       

                                  程序运行结果

代码:

#ifndef Plate_h
#define Plate_h

#include <string.h>
#include <vector>

#include <cv.h>
#include <highgui.h>
#include <cvaux.h>

using namespace std;
using namespace cv;

//车牌类
class Plate
    public:
        Plate();
        Plate(Mat img, Rect pos);
        string str();
        Rect position;//当前车牌在大图的位置,为了把识别出的车牌号显示到原图的车牌位置处
        Mat plateImg;//车牌图像,必须是灰度图像
        vector<char> chars;
        vector<Rect> charsPos;        
;

#endif
#include "stdafx.h"
#include "Plate.h"

Plate::Plate()


Plate::Plate(Mat img, Rect pos)
    plateImg=img;
    position=pos;

//将车牌号码按照间隔长短拼接成字符串
string Plate::str()
    string result="";
    //Order numbers
    vector<int> orderIndex;
    vector<int> xpositions;
    for(int i=0; i< charsPos.size(); i++)
        orderIndex.push_back(i);
        xpositions.push_back(charsPos[i].x);
    
    float min=xpositions[0];
    int minIdx=0;
    for(int i=0; i< xpositions.size(); i++)
        min=xpositions[i];
        minIdx=i;
        for(int j=i; j<xpositions.size(); j++)
            if(xpositions[j]<min)
                min=xpositions[j];
                minIdx=j;
            
        
        int aux_i=orderIndex[i];
        int aux_min=orderIndex[minIdx];
        orderIndex[i]=aux_min;
        orderIndex[minIdx]=aux_i;
        
        float aux_xi=xpositions[i];
        float aux_xmin=xpositions[minIdx];
        xpositions[i]=aux_xmin;
        xpositions[minIdx]=aux_xi;
    
    for(int i=0; i<orderIndex.size(); i++)
        result=result+chars[orderIndex[i]];
    
    return result;

                                                      车牌类代码

#ifndef OCR_h
#define OCR_h

#include <string.h>
#include <vector>

#include "Plate.h"

#include <cv.h>
#include <highgui.h>
#include <cvaux.h>
#include <ml.h>

using namespace std;
using namespace cv;


#define HORIZONTAL    1
#define VERTICAL    0

class CharSegment
public:
    CharSegment();
    CharSegment(Mat i, Rect p);
    Mat img;
    Rect pos;
;

class OCR
    public:
        bool DEBUG;
        bool saveSegments;
        string filename;
        static const int numCharacters;//字符个数
        static const char strCharacters[];//字符数组
        OCR(string trainFile);
        OCR();
        string run(Plate *input);//识别车牌
        int charSize;
        Mat preprocessChar(Mat in);//将字符图片调整为正方形
        int classify(Mat f);//根据特征识别出每个字符图片的字符
        void train(Mat trainData, Mat trainClasses, int nlayers);//训练分类器
        int classifyKnn(Mat f);//扩展的Knn分类器
        void trainKnn(Mat trainSamples, Mat trainClasses, int k);
        Mat features(Mat input, int size);//提取每幅字符图片的特征

    private:
        bool trained;
        vector<CharSegment> segment(Plate input);//分割车片图片
        Mat Preprocess(Mat in, int newSize);//缩放为正方形       
        Mat getVisualHistogram(Mat *hist, int type);//生成视觉直方图
        void drawVisualFeatures(Mat character, Mat hhist, Mat vhist, Mat lowData);//绘制视觉直方图
        Mat ProjectedHistogram(Mat img, int t);//计算累计直方图
        bool verifySizes(Mat r);//判断字符图像大小是否合适
        CvANN_MLP  ann;//神经网络分类器
        CvKNearest knnClassifier;//扩展的k邻域分类器
        int K;
;

#endif
#include "stdafx.h"
#include "OCR.h"

const char OCR::strCharacters[] = '0','1','2','3','4','5','6','7','8','9','B', 'C', 'D', 'F', 'G', 'H', 'J', 'K', 'L', 'M', 'N', 'P', 'R', 'S', 'T', 'V', 'W', 'X', 'Y', 'Z';
const int OCR::numCharacters=30;

CharSegment::CharSegment()
CharSegment::CharSegment(Mat i, Rect p)
    img=i;
    pos=p;


OCR::OCR()
    DEBUG=false;
    trained=false;
    saveSegments=false;
    charSize=20;

OCR::OCR(string trainFile)
    DEBUG=false;
    trained=false;
    saveSegments=false;
    charSize=20;

    //Read file storage.
    FileStorage fs;
    fs.open("OCR.xml", FileStorage::READ);
    Mat TrainingData;
    Mat Classes;
    fs["TrainingDataF15"] >> TrainingData;
    fs["classes"] >> Classes;

    train(TrainingData, Classes, 10);


//将单个字符图像变成正方形
Mat OCR::preprocessChar(Mat in)
    int h=in.rows;
    int w=in.cols;
    Mat transformMat=Mat::eye(2,3,CV_32F);//缩放矩阵
    int m=max(w,h);
    transformMat.at<float>(0,2)=m/2 - w/2;
    transformMat.at<float>(1,2)=m/2 - h/2;

    Mat warpImage(m,m, in.type());
    warpAffine(in, warpImage, transformMat, warpImage.size(), INTER_LINEAR, BORDER_CONSTANT, Scalar(0) );

    Mat out;
    resize(warpImage, out, Size(charSize, charSize) ); 

    return out;


//判断字符图像长宽是否符合要求
bool OCR::verifySizes(Mat r)
    //Char sizes 45x77
    float aspect=45.0f/77.0f;
    float charAspect= (float)r.cols/(float)r.rows;
    float error=0.35;
    float minHeight=15;
    float maxHeight=28;
    //We have a different aspect ratio for number 1, and it can be ~0.2
    float minAspect=0.2;
    float maxAspect=aspect+aspect*error;
    //area of pixels
    float area=countNonZero(r);
    //bb area
    float bbArea=r.cols*r.rows;
    //% of pixel in area
    float percPixels=area/bbArea;

    if(DEBUG)
        cout << "Aspect: "<< aspect << " ["<< minAspect << "," << maxAspect << "] "  << "Area "<< percPixels <<" Char aspect " << charAspect  << " Height char "<< r.rows << "\\n";
    if(percPixels < 0.8 && charAspect > minAspect && charAspect < maxAspect && r.rows >= minHeight && r.rows < maxHeight)
        return true;
    else
        return false;



//将车牌图像进一步分割成单个字符图片
vector<CharSegment> OCR::segment(Plate plate)
    Mat input=plate.plateImg;
    vector<CharSegment> output;
    //将输入图像二值化
    Mat img_threshold;
    threshold(input, img_threshold, 60, 255, CV_THRESH_BINARY_INV);
    if(DEBUG)
        imshow("Threshold plate", img_threshold);
    Mat img_contours;
    img_threshold.copyTo(img_contours);
    //查找字符轮廓
    vector< vector< Point> > contours;
    findContours(img_contours,
            contours, // 轮廓
            CV_RETR_EXTERNAL, //去除内环
            CV_CHAIN_APPROX_NONE); // 轮廓所有像素
    
    //将轮廓绘制到车牌图
    cv::Mat result;
    img_threshold.copyTo(result);
    cvtColor(result, result, CV_GRAY2RGB);
    cv::drawContours(result,contours,-1,cv::Scalar(255,0,0),1); 

    vector<vector<Point> >::iterator itc= contours.begin();
    
    //筛选符合条件的闭环   
    while (itc!=contours.end()) 
        
        //创建一个包围矩形
        Rect mr= boundingRect(Mat(*itc));
        rectangle(result, mr, Scalar(0,255,0));
        
        Mat auxRoi(img_threshold, mr);
        if(verifySizes(auxRoi))//判断长宽是否满足
            auxRoi=preprocessChar(auxRoi);//缩放成正方形
            output.push_back(CharSegment(auxRoi, mr));//保存字符图像及位置
            rectangle(result, mr, Scalar(0,125,255));
        
        ++itc;
    
    if(DEBUG)
        cout << "Num chars: " << output.size() << "\\n";
    if(DEBUG)
        imshow("SEgmented Chars", result);
    return output;

//计算累计直方图,统计每列或行的非0像素个数
Mat OCR::ProjectedHistogram(Mat img,int t)

    int sz=(t)?img.rows:img.cols;
    Mat mhist=Mat::zeros(1,sz,CV_32F);

    for(int j=0; j<sz; j++)
        Mat data=(t)?img.row(j):img.col(j);
        mhist.at<float>(j)=countNonZero(data);//
    

    //直方图归1化
    double min, max;
    minMaxLoc(mhist, &min, &max);
    
    if(max>0)
        mhist.convertTo(mhist,-1 , 1.0f/max, 0);

    return mhist;


//得到直方图图像
Mat OCR::getVisualHistogram(Mat *hist, int type)


    int size=100;
    Mat imHist;


    if(type==HORIZONTAL)
        imHist.create(Size(size,hist->cols), CV_8UC3);
    else
        imHist.create(Size(hist->cols, size), CV_8UC3);
    

    imHist=Scalar(55,55,55);

    for(int i=0;i<hist->cols;i++)
        float value=hist->at<float>(i);
        int maxval=(int)(value*size);

        Point pt1;
        Point pt2, pt3, pt4;

        if(type==HORIZONTAL)
            pt1.x=pt3.x=0;
            pt2.x=pt4.x=maxval;
            pt1.y=pt2.y=i;
            pt3.y=pt4.y=i+1;

            line(imHist, pt1, pt2, CV_RGB(220,220,220),1,8,0);
            line(imHist, pt3, pt4, CV_RGB(34,34,34),1,8,0);

            pt3.y=pt4.y=i+2;
            line(imHist, pt3, pt4, CV_RGB(44,44,44),1,8,0);
            pt3.y=pt4.y=i+3;
            line(imHist, pt3, pt4, CV_RGB(50,50,50),1,8,0);
        else

                        pt1.x=pt2.x=i;
                        pt3.x=pt4.x=i+1;
                        pt1.y=pt3.y=100;
                        pt2.y=pt4.y=100-maxval;


            line(imHist, pt1, pt2, CV_RGB(220,220,220),1,8,0);
            line(imHist, pt3, pt4, CV_RGB(34,34,34),1,8,0);

            pt3.x=pt4.x=i+2;
            line(imHist, pt3, pt4, CV_RGB(44,44,44),1,8,0);
            pt3.x=pt4.x=i+3;
            line(imHist, pt3, pt4, CV_RGB(50,50,50),1,8,0);
        
    
    return imHist ;


void OCR::drawVisualFeatures(Mat character, Mat hhist, Mat vhist, Mat lowData)
    Mat img(121, 121, CV_8UC3, Scalar(0,0,0));
    Mat ch;
    Mat ld;
    
    cvtColor(character, ch, CV_GRAY2RGB);

    resize(lowData, ld, Size(100, 100), 0, 0, INTER_NEAREST );
    cvtColor(ld,ld,CV_GRAY2RGB);

    Mat hh=getVisualHistogram(&hhist, HORIZONTAL);
    Mat hv=getVisualHistogram(&vhist, VERTICAL);

    Mat subImg=img(Rect(0,101,20,20));
    ch.copyTo(subImg);

    subImg=img(Rect(21,101,100,20));
    hh.copyTo(subImg);

    subImg=img(Rect(0,0,20,100));
    hv.copyTo(subImg);

    subImg=img(Rect(21,0,100,100));
    ld.copyTo(subImg);

    line(img, Point(0,100), Point(121,100), Scalar(0,0,255));
    line(img, Point(20,0), Point(20,121), Scalar(0,0,255));

    imshow("Visual Features", img);

    cvWaitKey(0);


Mat OCR::features(Mat in, int sizeData)
    //分别获取垂直和水平直方图信息
    Mat vhist=ProjectedHistogram(in,VERTICAL);
    Mat hhist=ProjectedHistogram(in,HORIZONTAL);
    
    //低分辨率图像
    Mat lowData;
    resize(in, lowData, Size(sizeData, sizeData) );//15x15

    if(DEBUG)
        drawVisualFeatures(in, hhist, vhist, lowData);
    
    //整合低分辨路图像信息和直方图统计信息,
    int numCols=vhist.cols+hhist.cols+lowData.cols*lowData.cols;
    
    Mat out=Mat::zeros(1,numCols,CV_32F);
    //保存特征信息
    int j=0;
    for(int i=0; i<vhist.cols; i++)
    
        out.at<float>(j)=vhist.at<float>(i);
        j++;
    
    for(int i=0; i<hhist.cols; i++)
    
        out.at<float>(j)=hhist.at<float>(i);
        j++;
    
    for(int x=0; x<lowData.cols; x++)
    
        for(int y=0; y<lowData.rows; y++)
            out.at<float>(j)=(float)lowData.at<unsigned char>(x,y);
            j++;
        
    
    if(DEBUG)
        cout << out << "\\n===========================================\\n";
    return out;


//训练        //训练样本数据//每条数据对应的字母下标//深度
void OCR::train(Mat TrainData, Mat classes, int nlayers)
    Mat layers(1,3,CV_32SC1);
    layers.at<int>(0)= TrainData.cols;//每个样本宽度
    layers.at<int>(1)= nlayers;//深度
    layers.at<int>(2)= numCharacters;//结果个数
    ann.create(layers, CvANN_MLP::SIGMOID_SYM, 1, 1);

    Mat trainClasses;
    trainClasses.create( TrainData.rows, numCharacters, CV_32FC1 );//每一条样本都对应着numCharacters个可能结果,但是只有一个结果是正确的,
    for( int i = 0; i <  trainClasses.rows; i++ )
    
        for( int k = 0; k < trainClasses.cols; k++ )
        
            //将该条训练数据对应的字符下标位置赋值为1,其他赋值为0
            if( k == classes.at<int>(i) )
                trainClasses.at<float>(i,k) = 1;
            else
                trainClasses.at<float>(i,k) = 0;
        
    
    Mat weights( 1, TrainData.rows, CV_32FC1, Scalar::all(1) );
    //开始训练学习
    ann.train( TrainData, trainClasses, weights );
    trained=true;

//识别字符
int OCR::classify(Mat f)
    int result=-1;
    Mat output(1, numCharacters, CV_32FC1);
    ann.predict(f, output);
    Point maxLoc;
    double maxVal;
    minMaxLoc(output, 0, &maxVal, 0, &maxLoc);//求最大值以及下标位置,这里没有打印出来最大值
    return maxLoc.x;


int OCR::classifyKnn(Mat f)
    int response = (int)knnClassifier.find_nearest( f, K );
    return response;

void OCR::trainKnn(Mat trainSamples, Mat trainClasses, int k)
    K=k;
    // learn classifier
    knnClassifier.train( trainSamples, trainClasses, Mat(), false, K );


string OCR::run(Plate *input)
    
    //分割车牌中每个字符
    vector<CharSegment> segments=segment(*input);

    for(int i=0; i<segments.size(); i++)
        //统一所有字符图像大小
        Mat ch=preprocessChar(segments[i].img);
        if(saveSegments)
            stringstream ss(stringstream::in | stringstream::out);
            ss << "tmpChars/" << filename << "_" << i << ".jpg";
            imwrite(ss.str(),ch);
        
        //提取每个字符图像特征
        Mat f=features(ch,15);
        //For each segment feature Classify
        int character=classify(f);
        input->chars.push_back(strCharacters[character]);
        input->charsPos.push_back(segments[i].pos);
    
    return "-";//input->str();



                                                                    OCR类代码

#include "stdafx.h"
#include <cv.h>
#include <highgui.h>
#include <cvaux.h>
#include <ml.h>

#include <iostream>
#include <vector>

#include "DetectRegions.h"
#include "OCR.h"

using namespace std;
using namespace cv;

string getFilename(string s) 

	char sep = '/';
	char sepExt = '.';

#ifdef _WIN32
	sep = '\\\\';
#endif

	size_t i = s.rfind(sep, s.length());
	if (i != string::npos) 
		string fn = (s.substr(i + 1, s.length() - i));
		size_t j = fn.rfind(sepExt, fn.length());
		if (i != string::npos) 
			return fn.substr(0, j);
		
		else
			return fn;
		
	
	else
		return "";
	


int main(int argc, char** argv)

	char* filename;
	Mat input_image;//必须为灰度图像

	//有输入图片才继续
	if (argc >= 2)
	
		filename = argv[1];
		input_image = imread(filename, 1);
	
	else
		printf("Use:\\n\\t%s image\\n", argv[0]);
		return 0;
	

	string filename_whithoutExt = getFilename(filename);//得到去除后缀部分
	OCR ocr("OCR.xml");//参数为保存了自己训练数据的xml文件
	ocr.saveSegments = true;
	ocr.DEBUG = true;
	ocr.filename = filename_whithoutExt;
	Plate plate;
	plate.plateImg = input_image;
	plate.position = Rect(50, 100, input_image.cols, input_image.rows);//车牌是从大图中抠图出来的,这里说明车牌的位置和大小
	imwrite("plateImg.jpg", plate.plateImg);
	string plateNumber = ocr.run(&plate);
	string licensePlate = plate.str();
	cout << "================================================\\n";
	cout << "License plate number: " << licensePlate << "\\n";
	cout << "================================================\\n";
	rectangle(input_image, plate.position, Scalar(0, 0, 200));
	putText(input_image, licensePlate, Point(plate.position.x, plate.position.y), CV_FONT_HERSHEY_SIMPLEX, 1, Scalar(0, 0, 200), 2);
	if (false)
		imshow("Plate Detected seg", plate.plateImg);
		cvWaitKey(0);
	

	imshow("Plate Detected", input_image);
	for (;;)
	
		int c;
		c = cvWaitKey(10);
		if ((char)c == 27)
			break;
	
	return 0;

                                                                      main函数代码

有关ANN神经网络分类器的原理及训练请参考我的另一篇文章:http://blog.csdn.net/xukaiwen_2016/article/details/53293465

最后:例子中没有实现对中文的识别,其实原理都是一样的,大家可以自己寻找中文车牌的图片进行分类器训练即可,代码几乎不用修改。

需要代码以及分类器训练数据xml文件的话,评论留下邮箱。




以上是关于OpenCV实现车牌识别,OCR分割,ANN神经网络的主要内容,如果未能解决你的问题,请参考以下文章

opencv 车牌字符分割 ANN网络识别字符

ann

基于SVM与人工神经网络的车牌识别系统

基于人工神经网络的数字字符识别系统demo:字符去噪分割

OpenCV+Python识别车牌和字符分割

OCR移动端车牌识别SDK:打造优质APP