开发环境

64 bits Windows OS (Win8.1)
VS2013
OpenCV 2.4.9

功能原理

算法要求

完成将Camera拍摄的手掌图片中分割出每个手指用于指纹识别

算法流程

核心代码

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

#ifdef TIME_RUN_COST

double duration = static_cast<double>(cv::getTickCount());//time

#endif

cout << "filename=" << filename <<endl;

Mat src = imread(filename, CV_LOAD_IMAGE_COLOR);

if (src.empty())

{

cout << "imread error!!!";

getchar();

return -1;

}

#ifdef BOB_DBG_COM

memset(out_filename, 0, sizeof(out_filename) / sizeof(char));

sprintf(out_filename, "%s-%s.jpg",out_name,"0-0src");

imwrite(out_filename, src);

#endif // BOB_DBG_COM

#if 0

int scaleSize = 4;

resize(src, src, Size(src.cols / scaleSize, src.rows / scaleSize), 0, 0, CV_INTER_AREA);

memset(out_filename, 0, sizeof(out_filename) / sizeof(char));

sprintf(out_filename, "%s-%s.jpg", out_name, "0-0src");

imwrite(out_filename, src);

#endif

#if 1

cout << "cut..." << endl;

int width = src.cols;

int height = src.rows;

float scale = 0.8;

cout << "width=" << width << ",height=" << height << endl;

Rect rect(0, 0, width, height*scale);

Mat imgCut;

imgCut = src(rect).clone();

#endif

//Mat imgCut = src;

cout << "filter..." << endl;

// filter2D(imgCut, imgCut, -1, kernel);

GaussianBlur(imgCut, imgCut, Size(5, 5), 0, 0);

// blur(imgCut, imgCut, Size(5, 5));

#if 0

cout << "EqualizeHist..." << endl;

Mat matOutEqualizeHist = Mat(imgCut.size(), CV_8UC3);

//IplImage* pImgOutEqualizeHist = cvCreateImage(cvSize(cameraFrame.cols, cameraFrame.rows), IPL_DEPTH_8U, 3);

IplImage pImgInEqualizeHist = (IplImage)(imgCut); // Mat-> IplImage

IplImage* pImgOutEqualizeHist = EqualizeHistColorImage(&pImgInEqualizeHist);

matOutEqualizeHist = pImgOutEqualizeHist; //IplImage -> Mat

#endif

// out

Mat imgSrc = Mat(imgCut.size(), CV_8UC1);

imgCut.copyTo(imgSrc);

Mat imgContour = Mat(imgSrc.size(), CV_8UC1);

#ifdef FINGER_EXTRACT_AT_NIGHT

cout << "Nigth,Threshold..." << endl;

Mat imgTmp;// = Mat(imgCut.size(), CV_8UC1);

cvtColor(imgSrc, imgTmp, CV_RGB2GRAY);

cvThresholdOtsu(&((IplImage)imgTmp), &((IplImage)imgTmp));

imgTmp.copyTo(imgContour);

#else

cout << "Day,Skin..." << endl;

Mat imgSkin2 = Mat(imgSrc.size(), CV_8UC1);

IplImage* pImgSkin2 = cvCreateImage(cvSize(imgSrc.cols, imgSrc.rows), IPL_DEPTH_8U, 1);

IplImage pImg2 = (IplImage)(imgSrc); // Mat-> IplImage

cvSkinOtsu(&pImg2, pImgSkin2);

imgSkin2 = pImgSkin2; //IplImage -> Mat

//Mat imgSkin = Mat(imgSrc.size(), CV_8UC1);

imgSkin2.copyTo(imgContour);

#endif

/// Contours

cout << "Find Contours..." << endl;

vector<vector<cv::Point> > contours;

vector<Vec4i> hierarchy;

findContours(imgContour, contours, hierarchy,

CV_RETR_EXTERNAL, CV_CHAIN_APPROX_SIMPLE, cv::Point(0, 0));

sort(contours.begin(), contours.end(), compareContourAreas);

int contours_num = contours.size();

cout << "contours_num=" << contours_num << endl;

#if 0

vector<vector<Point>>::const_iterator itContours = contours.begin();

//for (int i = 0; i < contours.size(); i++)

for (; itContours != contours.end(); ++itContours)

{

cout << "Size: " << itContours->size() << endl;//每个轮廓包含的点数

}

#endif

#if 1 //usd

// Eliminate too short or too long contours

int cmin = 100; // minimum contour length

//int cmax= 1000; // maximum contour length

vector<vector<Point>>::const_iterator itc = contours.begin();

while (itc != contours.end())

{

//if (itc->size() < cmin || itc->size() > cmax)

if (itc->size() < cmin) {

itc = contours.erase(itc);

}

else

++itc;

}

contours_num = contours.size();

cout << endl << "contours_num after Eliminate=" << contours_num << endl;

#endif

// extract the contour img

cout << "Extract Contours..." << endl;

if (contours_num >= 4)

{

Mat img1, img2, img3, img4;

std::vector<cv::Point> biggest1Contour = contours[contours_num - 1];

std::vector<cv::Point> biggest2Contour = contours[contours_num - 2];

std::vector<cv::Point> biggest3Contour = contours[contours_num - 3];

std::vector<cv::Point> biggest4Contour = contours[contours_num - 4];

std::vector<cv::Point> smallestContour = contours[0];

extractFingerImg2(contours, imgSrc, img1, contours_num, 1);

extractFingerImg2(contours, imgSrc, img2, contours_num, 2);

extractFingerImg2(contours, imgSrc, img3, contours_num, 3);

extractFingerImg2(contours, imgSrc, img4, contours_num, 4);

}

else if (contours_num == 3)

{

Mat img1, img2, img3;

std::vector<cv::Point> biggest1Contour = contours[contours_num - 1];

std::vector<cv::Point> biggest2Contour = contours[contours_num - 2];

std::vector<cv::Point> biggest3Contour = contours[contours_num - 3];

std::vector<cv::Point> smallestContour = contours[0];

extractFingerImg2(contours, imgSrc, img1, contours_num, 1);

extractFingerImg2(contours, imgSrc, img2, contours_num, 2);

extractFingerImg2(contours, imgSrc, img3, contours_num, 3);

}

else if (contours_num == 2)

{

Mat img1, img2;

std::vector<cv::Point> biggest1Contour = contours[contours_num - 1];

std::vector<cv::Point> biggest2Contour = contours[contours_num - 2];

std::vector<cv::Point> smallestContour = contours[0];

extractFingerImg2(contours, imgSrc, img1, contours_num, 1);

extractFingerImg2(contours, imgSrc, img2, contours_num, 2);

}

else if (contours_num == 1)

{

Mat img1;

std::vector<cv::Point> biggest1Contour = contours[contours_num - 1];

std::vector<cv::Point> smallestContour = contours[0];

extractFingerImg2(contours, imgSrc, img1, contours_num, 1);

}

else

{

cout << "error" << endl;

}

#ifdef TIME_RUN_COST

duration = static_cast<double>(cv::getTickCount()) - duration;

duration /= cv::getTickFrequency(); // the elapsed time in ms

cout << "time cost=" << duration << "s"<<endl;

#endif

#ifdef BOB_DBG_COM

memset(out_filename, 0, sizeof(out_filename) / sizeof(char));

sprintf(out_filename, "%s-%s.jpg", out_name, "4-imgContoursInSrc");

imwrite(out_filename, imgSrc);

//imwrite("4-imgContoursInSrc.jpg", imgSrc);

#endif // BOB_DBG_COM

算法效果

白天复杂场景

晚上场景

批量测试场景

转至：http://skyseraph.com/2014/07/24/CV/%E6%89%8B%E6%8E%8C%E6%89%8B%E6%8C%87%E5%88%86%E5%89%B2%E7%AE%97%E6%B3%95/

转载于:https://www.cnblogs.com/Ph-one/p/7364976.html

手掌手指分割算法（源码）相关推荐

超像素SLIC算法源码阅读
超像素SLIC算法源码阅读超像素SLIC算法源码阅读 SLIC简介源码阅读实验结果其他超像素算法对比超像素SLIC算法源码阅读 SLIC简介 SLIC的全称Simple Linear Ite ...
[unity3d]recast navigation navmesh 导航网格寻路算法源码分析
recast navigation navmesh导航网格算法源码分析 Author: 林绍川 recast navigation navmesh是unity3d ue4内置的寻路算法本文为了方便 ...
AidLux“实时头发分割”案例源码详解
"实时头发分割"案例源码详解 1. 构建APP框架和添加主要控件 2. 头发分割的方法打开实时头发分割案例 1.在VScode中进入代码编辑状态. 2.代码存在路径在/examp ...
基于Python Unet的医学影像分割系统源码，含皮肤病的数据及皮肤病分割的模型，用户输入图像，模型可以自动分割去皮肤病的区域
手把手教你用Unet做医学图像分割我们用Unet来做医学图像分割.我们将会以皮肤病的数据作为示范,训练一个皮肤病分割的模型出来,用户输入图像,模型可以自动分割去皮肤病的区域和正常的区域.废话不多说, ...
基于新唐M0的XXTEA加密解密算法源码
源:基于新唐M0的XXTEA加密解密算法源码 /*--------------------------------------------------------------------------- ...
[转] GIS算法源码集合
其他GIS相关代码下载索引 http://www.mygis.com.cn/codeindex10.htm 1．深度优先实现的路径分析源码 http://www.mygis.com.cn/codes/ ...
机器学习算法源码全解析（三）-范数规则化之核范数与规则项参数选择
前言参见上一篇博文,我们聊到了L0,L1和L2范数,这篇我们絮叨絮叨下核范数和规则项参数选择.知识有限,以下都是我一些浅显的看法,如果理解存在错误,希望大家不吝指正.谢谢. 机器学习算法源码全解析( ...
Learning to Rank中Pointwise关于PRank算法源码实现
[学习排序] Learning to Rank中Pointwise关于PRank算法源码实现标签: 学习排序PRankPointwiseLearning to Rank代码实现 2015-01-28 ...
c iostream.源码_通达信常用指标：「金手指」指标源码
通达信常用指标:[金手指 ]指标源码强弱:=IF(C>=MA(C,18) AND (MA(C,18)>=REF(MA(C,18),1) OR (MA(C,18)<REF(MA(C, ...

手掌手指分割算法（源码）

开发环境

功能原理

算法效果

手掌手指分割算法（源码）相关推荐

最新文章

热门文章