|
|
#include "../include/chars_segment.h"
|
|
|
|
|
|
/*! \namespace easypr
|
|
|
Namespace where all the C++ EasyPR functionality resides
|
|
|
*/
|
|
|
namespace easypr{
|
|
|
|
|
|
const float DEFAULT_BLUEPERCEMT = 0.3;
|
|
|
const float DEFAULT_WHITEPERCEMT = 0.1;
|
|
|
|
|
|
CCharsSegment::CCharsSegment()
|
|
|
{
|
|
|
//cout << "CCharsSegment" << endl;
|
|
|
m_LiuDingSize = DEFAULT_LIUDING_SIZE;
|
|
|
m_theMatWidth = DEFAULT_MAT_WIDTH;
|
|
|
|
|
|
//！车牌颜色判断参数
|
|
|
m_ColorThreshold = DEFAULT_COLORTHRESHOLD;
|
|
|
m_BluePercent = DEFAULT_BLUEPERCEMT;
|
|
|
m_WhitePercent = DEFAULT_WHITEPERCEMT;
|
|
|
|
|
|
m_debug = DEFAULT_DEBUG;
|
|
|
}
|
|
|
|
|
|
//! 字符尺寸验证
|
|
|
bool CCharsSegment::verifySizes(Mat r){
|
|
|
//Char sizes 45x90
|
|
|
float aspect=45.0f/90.0f;
|
|
|
float charAspect= (float)r.cols/(float)r.rows;
|
|
|
float error=0.7;
|
|
|
float minHeight=10;
|
|
|
float maxHeight=35;
|
|
|
//We have a different aspect ratio for number 1, and it can be ~0.2
|
|
|
float minAspect=0.05;
|
|
|
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(percPixels <= 1 && charAspect > minAspect && charAspect < maxAspect && r.rows >= minHeight && r.rows < maxHeight)
|
|
|
return true;
|
|
|
else
|
|
|
return false;
|
|
|
}
|
|
|
|
|
|
//! 字符预处理
|
|
|
Mat CCharsSegment::preprocessChar(Mat in){
|
|
|
//Remap image
|
|
|
int h=in.rows;
|
|
|
int w=in.cols;
|
|
|
int charSize=CHAR_SIZE; //统一每个字符的大小
|
|
|
Mat transformMat=Mat::eye(2,3,CV_32F);
|
|
|
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;
|
|
|
}
|
|
|
|
|
|
//! 直方图均衡，为判断车牌颜色做准备
|
|
|
Mat CCharsSegment::histeq(Mat in)
|
|
|
{
|
|
|
Mat out(in.size(), in.type());
|
|
|
if(in.channels()==3)
|
|
|
{
|
|
|
Mat hsv;
|
|
|
vector<Mat> hsvSplit;
|
|
|
cvtColor(in, hsv, CV_BGR2HSV);
|
|
|
split(hsv, hsvSplit);
|
|
|
equalizeHist(hsvSplit[2], hsvSplit[2]);
|
|
|
merge(hsvSplit, hsv);
|
|
|
cvtColor(hsv, out, CV_HSV2BGR);
|
|
|
}
|
|
|
else if(in.channels()==1)
|
|
|
{
|
|
|
equalizeHist(in, out);
|
|
|
}
|
|
|
return out;
|
|
|
}
|
|
|
|
|
|
//getPlateType
|
|
|
//判断车牌的类型，1为蓝牌，2为黄牌，0为未知，默认蓝牌
|
|
|
//通过像素中蓝色所占比例的多少来判断，大于0.3为蓝牌，否则为黄牌
|
|
|
int CCharsSegment::getPlateType(Mat input)
|
|
|
{
|
|
|
Mat img;
|
|
|
input.copyTo(img);
|
|
|
img = histeq(img);
|
|
|
|
|
|
double countBlue = 0;
|
|
|
double countWhite = 0;
|
|
|
|
|
|
int nums = img.rows*img.cols;
|
|
|
for(int i=0; i < img.rows; i++)
|
|
|
{
|
|
|
for(int j=0; j < img.cols; j++)
|
|
|
{
|
|
|
Vec3b intensity = img.at<Vec3b>(i,j);
|
|
|
int blue = int(intensity.val[0]);
|
|
|
int green = int(intensity.val[1]);
|
|
|
int red = int(intensity.val[2]);
|
|
|
|
|
|
if(blue > m_ColorThreshold && green > 10 && red > 10)
|
|
|
countBlue++;
|
|
|
|
|
|
if(blue > m_ColorThreshold && green > m_ColorThreshold && red > m_ColorThreshold)
|
|
|
countWhite++;
|
|
|
}
|
|
|
}
|
|
|
|
|
|
double percentBlue = countBlue/nums;
|
|
|
double percentWhite = countWhite/nums;
|
|
|
|
|
|
if (percentBlue - m_BluePercent > 0 && percentWhite - m_WhitePercent > 0)
|
|
|
return 1;
|
|
|
else
|
|
|
return 2;
|
|
|
|
|
|
return 0;
|
|
|
}
|
|
|
|
|
|
//clearLiuDing
|
|
|
//去除车牌上方的钮钉
|
|
|
//计算每行元素的阶跃数，如果小于X认为是柳丁，将此行全部填0（涂黑）
|
|
|
//X的推荐值为，可根据实际调整
|
|
|
Mat CCharsSegment::clearLiuDing(Mat img)
|
|
|
{
|
|
|
const int x = m_LiuDingSize;
|
|
|
Mat jump = Mat::zeros(1, img.rows, CV_32F);
|
|
|
for(int i=0; i < img.rows; i++)
|
|
|
{
|
|
|
int jumpCount = 0;
|
|
|
for(int j=0; j < img.cols-1; j++)
|
|
|
{
|
|
|
if (img.at<char>(i,j) != img.at<char>(i,j+1))
|
|
|
jumpCount++;
|
|
|
}
|
|
|
jump.at<float>(i) = jumpCount;
|
|
|
}
|
|
|
for(int i=0; i < img.rows; i++)
|
|
|
{
|
|
|
if(jump.at<float>(i) <= x)
|
|
|
{
|
|
|
for(int j=0; j < img.cols; j++)
|
|
|
{
|
|
|
img.at<char>(i,j) = 0;
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
return img;
|
|
|
}
|
|
|
|
|
|
//! 字符分割与排序
|
|
|
int CCharsSegment::charsSegment(Mat input, vector<Mat>& resultVec)
|
|
|
{
|
|
|
if( !input.data )
|
|
|
{ return -3; }
|
|
|
|
|
|
//判断车牌颜色以此确认threshold方法
|
|
|
int plateType = getPlateType(input);
|
|
|
cvtColor(input, input, CV_RGB2GRAY);
|
|
|
|
|
|
//Threshold input image
|
|
|
Mat img_threshold;
|
|
|
if (1 == plateType)
|
|
|
threshold(input, img_threshold, 10, 255, CV_THRESH_OTSU+CV_THRESH_BINARY);
|
|
|
else
|
|
|
threshold(input, img_threshold, 10, 255, CV_THRESH_OTSU+CV_THRESH_BINARY_INV);
|
|
|
|
|
|
if(m_debug)
|
|
|
{
|
|
|
stringstream ss(stringstream::in | stringstream::out);
|
|
|
ss << "image/tmp/debug_char_threshold" << ".jpg";
|
|
|
imwrite(ss.str(), img_threshold);
|
|
|
}
|
|
|
|
|
|
//去除车牌上方的柳钉以及下方的横线等干扰
|
|
|
clearLiuDing(img_threshold);
|
|
|
|
|
|
|
|
|
if(m_debug)
|
|
|
{
|
|
|
stringstream ss(stringstream::in | stringstream::out);
|
|
|
ss << "image/tmp/debug_char_clearLiuDing" << ".jpg";
|
|
|
imwrite(ss.str(), img_threshold);
|
|
|
}
|
|
|
|
|
|
Mat img_contours;
|
|
|
img_threshold.copyTo(img_contours);
|
|
|
|
|
|
vector< vector< Point> > contours;
|
|
|
findContours(img_contours,
|
|
|
contours, // a vector of contours
|
|
|
CV_RETR_EXTERNAL, // retrieve the external contours
|
|
|
CV_CHAIN_APPROX_NONE); // all pixels of each contours
|
|
|
|
|
|
//Start to iterate to each contour founded
|
|
|
vector<vector<Point> >::iterator itc= contours.begin();
|
|
|
|
|
|
vector<Rect> vecRect;
|
|
|
|
|
|
//Remove patch that are no inside limits of aspect ratio and area.
|
|
|
//将不符合特定尺寸的图块排除出去
|
|
|
while (itc != contours.end())
|
|
|
{
|
|
|
Rect mr = boundingRect(Mat(*itc));
|
|
|
Mat auxRoi(img_threshold, mr);
|
|
|
if (verifySizes(auxRoi))
|
|
|
{
|
|
|
vecRect.push_back(mr);
|
|
|
}
|
|
|
++itc;
|
|
|
}
|
|
|
|
|
|
if (vecRect.size() == 0)
|
|
|
return -3;
|
|
|
|
|
|
vector<Rect> sortedRect(vecRect);
|
|
|
//对符合尺寸的图块按照从左到右进行排序
|
|
|
std::sort(sortedRect.begin(), sortedRect.end(), [] (const Rect &r1, const Rect &r2) {
|
|
|
return r1.x < r2.x;
|
|
|
});
|
|
|
|
|
|
int specIndex = 0;
|
|
|
//获得指示城市的特定Rect,如苏A的"A"
|
|
|
specIndex = GetSpecificRect(sortedRect);
|
|
|
|
|
|
if(m_debug)
|
|
|
{
|
|
|
if (specIndex < sortedRect.size())
|
|
|
{
|
|
|
Mat specMat(img_threshold, sortedRect[specIndex]);
|
|
|
stringstream ss(stringstream::in | stringstream::out);
|
|
|
ss << "image/tmp/debug_specMat" <<".jpg";
|
|
|
imwrite(ss.str(), specMat);
|
|
|
}
|
|
|
}
|
|
|
|
|
|
//根据特定Rect向左反推出中文字符
|
|
|
//这样做的主要原因是根据findContours方法很难捕捉到中文字符的准确Rect，因此仅能
|
|
|
//退过特定算法来指定
|
|
|
Rect chineseRect;
|
|
|
if (specIndex < sortedRect.size())
|
|
|
chineseRect = GetChineseRect(sortedRect[specIndex]);
|
|
|
else
|
|
|
return -3;
|
|
|
|
|
|
if(m_debug)
|
|
|
{
|
|
|
Mat chineseMat(img_threshold, chineseRect);
|
|
|
stringstream ss(stringstream::in | stringstream::out);
|
|
|
ss << "image/tmp/debug_chineseMat" <<".jpg";
|
|
|
imwrite(ss.str(), chineseMat);
|
|
|
}
|
|
|
|
|
|
|
|
|
//新建一个全新的排序Rect
|
|
|
//将中文字符Rect第一个加进来，因为它肯定是最左边的
|
|
|
//其余的Rect只按照顺序去6个，车牌只可能是7个字符！这样可以避免阴影导致的“1”字符
|
|
|
vector<Rect> newSortedRect;
|
|
|
newSortedRect.push_back(chineseRect);
|
|
|
RebuildRect(sortedRect, newSortedRect, specIndex);
|
|
|
|
|
|
if (newSortedRect.size() == 0)
|
|
|
return -3;
|
|
|
|
|
|
for (int i = 0; i < newSortedRect.size(); i++)
|
|
|
{
|
|
|
Rect mr = newSortedRect[i];
|
|
|
Mat auxRoi(img_threshold, mr);
|
|
|
|
|
|
if (1)
|
|
|
{
|
|
|
auxRoi = preprocessChar(auxRoi);
|
|
|
if(m_debug)
|
|
|
{
|
|
|
stringstream ss(stringstream::in | stringstream::out);
|
|
|
ss << "image/tmp/debug_char_auxRoi_" << i <<".jpg";
|
|
|
imwrite(ss.str(), auxRoi);
|
|
|
}
|
|
|
resultVec.push_back(auxRoi);
|
|
|
}
|
|
|
}
|
|
|
|
|
|
return 0;
|
|
|
}
|
|
|
|
|
|
//! 根据特殊车牌来构造猜测中文字符的位置和大小
|
|
|
Rect CCharsSegment::GetChineseRect(const Rect rectSpe)
|
|
|
{
|
|
|
int height = rectSpe.height;
|
|
|
float newwidth = rectSpe.width * 1.15;
|
|
|
int x = rectSpe.x;
|
|
|
int y = rectSpe.y;
|
|
|
|
|
|
int newx = x - int (newwidth * 1.15);
|
|
|
newx = newx > 0 ? newx : 0;
|
|
|
|
|
|
Rect a(newx, y, int(newwidth), height);
|
|
|
|
|
|
return a;
|
|
|
}
|
|
|
|
|
|
//! 找出指示城市的字符的Rect，例如苏A7003X，就是"A"的位置
|
|
|
int CCharsSegment::GetSpecificRect(const vector<Rect>& vecRect)
|
|
|
{
|
|
|
vector<int> xpositions;
|
|
|
int maxHeight = 0;
|
|
|
int maxWidth = 0;
|
|
|
|
|
|
for (int i = 0; i < vecRect.size(); i++)
|
|
|
{
|
|
|
xpositions.push_back(vecRect[i].x);
|
|
|
|
|
|
if (vecRect[i].height > maxHeight)
|
|
|
{
|
|
|
maxHeight = vecRect[i].height;
|
|
|
}
|
|
|
if (vecRect[i].width > maxWidth)
|
|
|
{
|
|
|
maxWidth = vecRect[i].width;
|
|
|
}
|
|
|
}
|
|
|
|
|
|
int specIndex = 0;
|
|
|
for (int i = 0; i < vecRect.size(); i++)
|
|
|
{
|
|
|
Rect mr = vecRect[i];
|
|
|
int midx = mr.x + mr.width/2;
|
|
|
|
|
|
//如果一个字符有一定的大小，并且在整个车牌的1/7到2/7之间，则是我们要找的特殊车牌
|
|
|
if ((mr.width > maxWidth * 0.8 || mr.height > maxHeight * 0.8) &&
|
|
|
(midx < int(m_theMatWidth / 7) * 2 && midx > int(m_theMatWidth / 7) * 1))
|
|
|
{
|
|
|
specIndex = i;
|
|
|
}
|
|
|
}
|
|
|
|
|
|
return specIndex;
|
|
|
}
|
|
|
|
|
|
//! 这个函数做两个事情
|
|
|
// 1.把特殊字符Rect左边的全部Rect去掉，后面再重建中文字符的位置。
|
|
|
// 2.从特殊字符Rect开始，依次选择6个Rect，多余的舍去。
|
|
|
int CCharsSegment::RebuildRect(const vector<Rect>& vecRect, vector<Rect>& outRect, int specIndex)
|
|
|
{
|
|
|
//最大只能有7个Rect,减去中文的就只有6个Rect
|
|
|
int count = 6;
|
|
|
|
|
|
for (int i = 0; i < vecRect.size(); i++)
|
|
|
{
|
|
|
//将特殊字符左边的Rect去掉，这个可能会去掉中文Rect，不过没关系，我们后面会重建。
|
|
|
if (i < specIndex)
|
|
|
continue;
|
|
|
|
|
|
outRect.push_back(vecRect[i]);
|
|
|
if (!--count)
|
|
|
break;
|
|
|
}
|
|
|
|
|
|
return 0;
|
|
|
}
|
|
|
|
|
|
} /*! \namespace easypr*/
|
