直方图反向投影：

即取直方图中的值，按直方图面积由大到小，对其对应的像素也由大到小赋予新值。即某种灰度值在图像中所占面积越大，其对应的像素的新值就越大；反之就越小。

void calcBackProject( const Mat* images, //源数组/图像；有相同的大小和深度(CV_8U或CV_32F)。int nimages,//源图像的数量const int* channels, //计算反投影的通道列表；通道数量必须与直方图维度相匹配。InputArray hist,//输入直方图。OutputArray backProject, //目标反向投影数组，与images[0]大小、深度相同的单通道数组。const float** ranges,//每个维度的直方图bin边界数组。double scale = 1, //输出反向投影的可选比例因子。bool uniform = true //直方图是否均匀的标志。);

示例：

获取测试图像中每个像素的hue数据 hi,j，并找到 hi,j 在hue直方图中的bin的位置

#include<opencv2/opencv.hpp>using namespace cv;Mat hue;int bins = 10;void histBackprojection(int, void*);int main(){ Mat src, hsv; src = imread("E:/image/image/shape.jpg"); if (src.empty()) { printf("can not load image \n"); return -1; } namedWindow("input", WINDOW_AUTOSIZE); imshow("input", src); cvtColor(src, hsv, COLOR_BGR2HSV); hue.create(hsv.size(), hsv.depth()); int ch[] = {0, 0}; //分离Hue通道:色相通道 mixChannels(&hsv, 1, &hue, 1, ch, 1); //创建Trackbar来输入bin的数目 createTrackbar("bins:", "input", &bins, 255, histBackprojection); histBackprojection(0, 0); waitKey(0); return 0;}void histBackprojection(int, void*){ MatND hist, backproj; int histSize = max(bins, 2); float range[] = {0, 255}; const float *ranges = {range}; //计算直方图 calcHist(&hue, 1, 0, Mat(), hist,1, &histSize, &ranges, true, false); //将直方图bin的数值归一化到0-255 normalize(hist, hist, 0, 255, NORM_MINMAX, -1, Mat()); //计算反向投影 calcBackProject(&hue, 1, 0, hist, backproj, &ranges, 1, true); namedWindow("backprogection", WINDOW_AUTOSIZE); imshow("backprogection", backproj); //显示直方图 int binsW = cvRound((double)500/histSize); Mat histImg = Mat::zeros(500, 500, CV_8UC3); RNG rng(123); for (int i = 0; i < bins; i++) { Scalar color = Scalar(rng.uniform(0,255),rng.uniform(0,255),rng.uniform(0,255)); rectangle(histImg, Point(i*binsW, 500), Point((i+1)*binsW, 500 - cvRound(hist.at<float>(i) * 500 / 255.0)), color, -1); } namedWindow("histogram", WINDOW_AUTOSIZE); imshow("histogram", histImg);}

以上这篇opencv3/C++ 直方图反向投影实例就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持。