OpenCV算法精讲教材源码和测试图片

#include<opencv2/core/core.hpp> #include<opencv2/highgui/highgui.hpp> #include<opencv2/imgproc/imgproc.hpp> using namespace cv; //非极大值抑制 Mat non_maximum_suppression_default(Mat dx, Mat dy) { //使用平方和开方的方式计算边缘强度 Mat edgeMag; cv::magnitude(dx, dy, edgeMag); //宽高 int rows = dx.rows; int cols = dy.cols; //边缘强度的非极大抑制 Mat edgeMag_nonMaxSup = Mat::zeros(dx.size(), dx.type()); for (int r = 1; r < rows-1; r++) { for (int c = 1; c < cols-1; c++) { float x = dx.at<float>(r, c); float y = dy.at<float>(r, c); //梯度方向 float angle = atan2f(y, x)/CV_PI*180; //当前位置的边缘强度 float mag = edgeMag.at<float>(r, c); // 左 / 右方向比较 if (abs(angle) < 22.5 || abs(angle) > 157.5) { float left = edgeMag.at<float>(r, c - 1); float right = edgeMag.at<float>(r, c + 1); if (mag > left && mag > right) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //左上 / 右下方向比较 if ((angle >= 22.5 && angle < 67.5) || (angle < -112.5 && angle >= 157.5)) { float leftTop = edgeMag.at<float>(r - 1, c - 1); float rightBottom = edgeMag.at<float>(r + 1, c + 1); if (mag > leftTop && mag > rightBottom) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //上 / 下方向比较 if ((angle >= 67.5 && angle <= 112.5) || (angle >= -112.5 && angle <= -67.5)) { float top = edgeMag.at<float>(r - 1, c); float bottom = edgeMag.at<float>(r + 1, c); if (mag > top && mag > bottom) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //右上 / 左下方向 if ((angle > 112.5 && angle <= 157.5) || (angle > -67.5 && angle <= -22.5)) { float rightTop = edgeMag.at<float>(r - 1, c + 1); float leftBottom = edgeMag.at<float>(r + 1, c - 1); if (mag > rightTop && mag > leftBottom) edgeMag_nonMaxSup.at<float>(r, c) = mag; } } } return edgeMag_nonMaxSup; } //非极大值抑制 ： 插值比较 Mat non_maximum_suppression_Inter(Mat dx, Mat dy) { //使用平方和开方的方式计算边缘强度 Mat edgeMag; cv::magnitude(dx, dy, edgeMag); //宽高 int rows = dx.rows; int cols = dy.cols; //边缘强度的非极大抑制 Mat edgeMag_nonMaxSup = Mat::zeros(dx.size(), dx.type()); for (int r = 1; r < rows-1; r++) { for (int c = 1; c < cols-1; c++) { float x = dx.at<float>(r, c); float y = dy.at<float>(r, c); if (x == 0 && y == 0) continue; float angle = atan2f(y, x) / CV_PI * 180; //领域内八个方向上的边缘强度 float leftTop = edgeMag.at<float>(r - 1, c - 1); float top = edgeMag.at<float>(r - 1, c); float rightBottom = edgeMag.at<float>(r + 1, c + 1); float right = edgeMag.at<float>(r, c+1); float rightTop = edgeMag.at<float>(r - 1, c + 1); float leftBottom = edgeMag.at<float>(r + 1, c - 1); float bottom = edgeMag.at<float>(r + 1, c); float left = edgeMag.at<float>(r, c - 1); float mag = edgeMag.at<float>(r, c); //左上方与上方的插值 右下方和下方的插值 if ((angle > 45 && angle <= 90) || (angle > -135 && angle <= -90)) { float ratio = x / y; float top = edgeMag.at<float>(r - 1, c); //插值 float leftTop_top = ratio*leftTop + (1 - ratio)*top; float rightBottom_bottom = ratio*rightBottom + (1 - ratio)*bottom; if (mag > leftTop_top && mag > rightBottom_bottom) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //右上方和上方的插值 左下方和下方的插值 if ((angle > 90 && angle <= 135) || (angle > -90 && angle <= -45)) { float ratio = abs(x / y); float rightTop_top = ratio*rightTop + (1 - ratio)*top; float leftBottom_bottom = ratio*leftBottom + (1 - ratio)*bottom; if (mag > rightTop_top && mag > leftBottom_bottom) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //左上方和左方的插值 右下方和右方的插值 if ((angle >= 0 && angle <= 45) || (angle > -180 && angle <= -135)) { float ratio = y / x; float rightBottom_right = ratio*rightBottom + (1 - ratio)*right; float leftTop_left = ratio*leftTop + (1 - ratio)*left; if (mag > rightBottom_right && mag > leftTop_left) edgeMag_nonMaxSup.at<float>(r, c) = mag; } //右上方和右方的插值 左下方和左方的插值 if ((angle > 135 && angle <= 180) || (angle > -45 && angle <= 0)) { float ratio = abs(y / x); float rightTop_right = ratio*rightTop + (1 - ratio)*right; float leftBottom_left = ratio*leftBottom + (1 - ratio)*left; if (mag > rightTop_right && mag > leftBottom_left) edgeMag_nonMaxSup.at<float>(r, c) = mag; } } } return edgeMag_nonMaxSup; } //确定一个点的坐标是否在图像范围内 bool checkInRange(int r, int c, int rows, int cols) { if (r >= 0 && r < rows && c>=0 && c < cols) return true; else return false; } //从确定边缘点出发，延长边缘 void trace(Mat edgeMag_nonMaxSup, Mat &edge, float lowerThresh, int r, int c, int rows,int cols) { if (edge.at<uchar>(r, c) == 0) { edge.at<uchar>(r, c) = 255; for (int i = -1; i <= 1; i++) { for (int j = -1; j <= 1; j++) { float mag = edgeMag_nonMaxSup.at<float>(r+i, c+j); if (checkInRange(r + i, c + j, rows, cols) && mag >= lowerThresh) trace(edgeMag_nonMaxSup, edge, lowerThresh, r+i, c+j, rows, cols); } } } } //双阈值的滞后阈值处理 Mat hysteresisThreshold(Mat edgeMag_nonMaxSup, float lowerThresh, float upperThresh) { //宽高 int rows = edgeMag_nonMaxSup.rows; int cols = edgeMag_nonMaxSup.cols; //最后的边缘输出图 Mat edge = Mat::zeros(Size(cols, rows), CV_8UC1); //滞后阈值处理 for (int r = 1; r < rows - 1; r++) { for (int c = 1; c < cols-1; c++) { float mag = edgeMag_nonMaxSup.at<float>(r, c); //大于高阈值，可作为确定边缘点被接受 //并以该点位起始点延长边缘 if(mag>=upperThresh) trace(edgeMag_nonMaxSup, edge, lowerThresh, r, c, rows, cols); //小于低阈值的直接被剔除 if (mag < lowerThresh) edge.at<uchar>(r, c) = 0; } } return edge; } int main(int argc,char*argv[]) { //输入图像 Mat image = imread(argv[1], CV_LOAD_IMAGE_GRAYSCALE); if (!image.data) return -1; //第一步：sobel 卷积，这里采用 OpenCV 封装的 Sobel 函数 Mat dx, dy; cv::Sobel(image, dx, CV_32FC1, 1, 0, 3); cv::Sobel(image, dy, CV_32FC1, 0, 1, 3); //边缘强度的灰度级显示 Mat edgeMag; cv::magnitude(dx, dy, edgeMag); Mat mag; edgeMag.convertTo(mag, CV_8UC1); cv::imshow("边缘强度的灰度级显示", mag); cv::imwrite("mag.jpg", mag); //第二步：非极大值抑制 Mat edgeMag_nonMaxSup = non_maximum_suppression_default(dx, dy); //显示非极大值抑制后的边缘强度图 Mat nonMaxSup; edgeMag_nonMaxSup.convertTo(nonMaxSup, CV_8UC1); cv::imshow("非极大值抑制", nonMaxSup); cv::imwrite("nonMaxSup.jpg", nonMaxSup); //第三步：双阈值的滞后阈值处理 float lowerThresh = 40; float upperThresh = 150; Mat edge = hysteresisThreshold(edgeMag_nonMaxSup, lowerThresh, upperThresh); imshow("CannyEdge", edge); cv::imwrite("CannyEdge.jpg", edge); //与单阈值处理进行比较 Mat lowerThreshEdge; cv::threshold(edgeMag_nonMaxSup, lowerThreshEdge, lowerThresh, 255, CV_THRESH_BINARY); cv::convertScaleAbs(lowerThreshEdge, lowerThreshEdge, 1.0, 0); cv::imshow("低阈值处理后的效果", lowerThreshEdge); cv::imwrite("lowerThreshEdge.jpg", lowerThreshEdge); Mat upperThreshEdge; cv::threshold(edgeMag_nonMaxSup, upperThreshEdge, upperThresh, 255, CV_THRESH_BINARY); cv::convertScaleAbs(upperThreshEdge, upperThreshEdge, 1.0, 0.0); cv::imshow("高阈值处理后的效果", upperThreshEdge); cv::imwrite("upperThreshEdge.jpg", upperThreshEdge); waitKey(0); return 0; }