OpenCV成长之路：图像直方图的应用

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

int main()

{

    Mat face=imread("../face.png");        // 样本

    Mat ImgSrc=imread("../img.png");    // 待检测的图像

    //图像降维

    colorReduce(face,face,32);

    colorReduce(ImgSrc,ImgSrc,32);

    // 计算颜色直方图

    const int channels[3]={0,1,2};

    const int histSize[3]={256,256,256};

    float hranges[2]={0,255};

    const float* ranges[3]={hranges,hranges,hranges};

    MatND hist;

    calcHist(&face,1,channels,Mat(),hist,3,histSize,ranges);

    // 直方图归一化

    normalize(hist,hist,1.0);

    // 直方图反向映射

    Mat result;

    calcBackProject(&ImgSrc,1,channels,hist,result,ranges,255);

    // 将结果进行阈值化

    threshold(result,result,255*(0.05),255,THRESH_BINARY);

return 0;

}

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

int main()

{

    Mat refImg=imread("../ref.png");

    Mat image1=imread("../image1.png");

    Mat image2=imread("../image2.png");

    ColorHistogram imgHist;

    //图像颜色降维

    refImg=imgHist.colorReduce(refImg,64);

    image1=imgHist.colorReduce(image1,64);

    image2=imgHist.colorReduce(image2,64);

    MatND refH=imgHist.getHistogram(refImg);

    MatND hist1=imgHist.getHistogram(image1);

    MatND hist2=imgHist.getHistogram(image2);

    double dist1,dist2;

    dist1=compareHist(refH,hist1,CV_COMP_BHATTACHARYYA);

    dist2=compareHist(refH,hist2,CV_COMP_BHATTACHARYYA);

    std::cout<<"dist1="<<dist1<<",dist2="<<dist2<<std::endl;

    return 0;

}