基于opencv像素点匹配的数字识别.rar

#include <opencv2\opencv.hpp> #include<opencv2\imgproc\imgproc.hpp> #include <iostream> #include <math.h> using namespace std; using namespace cv; //创建记录轮廓数量和坐标的结构体 struct num_contours { double x, y; //轮廓的x、y坐标 int index; //轮廓的顺序 bool operator <(num_contours &m) { if (y > m.y+50) return false; else if (y > m.y-50) { if (x < m.x) return true; else return false; } else return true; } }num_contours[100]; //创建记录结果的结构体 struct result { double data; int num; bool operator<(result &m) { if (data < m.data)return true; else return false; } }result[100]; void select_result(Mat & src, int num); double compare(Mat &src, Mat &model); vector<Mat> imread_model(); void deal(Mat &src, Mat &model, int m); void select_result(Mat & src, int num); int main() { Mat src, gray_src, dest, dst; src = imread("D:/Project/Opencv-Project/num_text/num_text_03.png"); //加载原图 cvtColor(src, gray_src, COLOR_BGR2GRAY); //对图像进行预处理 threshold(gray_src, dest, 150, 255, THRESH_BINARY_INV); vector<vector<Point>> contours; vector<Vec4i> hierarchy; findContours(dest, contours, hierarchy, RETR_EXTERNAL, CHAIN_APPROX_NONE, Point(0, 0)); //提取轮廓 vector<Rect> rect; Rect min_bounding_rect,re_rect; float tl_x, br_y, width, height; for (int i = 0; i < contours.size(); i++) { if (contourArea(contours[i]) > 50) { min_bounding_rect = boundingRect(contours[i]); rect.push_back(boundingRect(contours[i])); tl_x = (float)min_bounding_rect.tl().x; br_y = (float)min_bounding_rect.br().y; width = (float)min_bounding_rect.width; height = (float)min_bounding_rect.height; num_contours[i].x = (tl_x * 2 + width) / 2.0; //计算轮廓中心点 num_contours[i].y = (br_y * 2 + height) / 2.0; num_contours[i].index = i; } } sort(num_contours, num_contours + contours.size()); //根据轮廓中心点坐标进行排序 vector<Mat> roi_rect(contours.size()); for (int j = 0; j < contours.size(); j++) //ROI处理分割对象 { int k = num_contours[j].index; //对分割对象进行预处理 src(rect[k]).copyTo(roi_rect[j]); cvtColor(roi_rect[j], roi_rect[j], COLOR_BGR2GRAY); threshold(roi_rect[j], roi_rect[j], 150, 255, THRESH_BINARY_INV); select_result(roi_rect[j], j + 1); } for (int i = 0; i < contours.size(); i++) { rectangle(src, rect[i], Scalar(0, 0, 255), 1, 8, 0); } imshow("二值图", dest); imshow("原图", src); waitKey(0); return 0; } double compare(Mat &src, Mat &model) //模板与图像对比计算相似度 { Mat re_model; resize(model, re_model, src.size()); int rows, cols; uchar *src_data, *model_data; rows = re_model.rows; cols = re_model.cols*src.channels(); double percentage,same=0.0,different=0.0; for (int i = 0; i < rows; i++) { src_data = src.ptr<uchar>(i); model_data = re_model.ptr<uchar>(i); for (int j = 0; j < cols; j++) { if (src_data[j] == model_data[j]) { same++; } else { different++; } } } percentage = same / (same + different); return percentage; //返回相似度 } vector<Mat> imread_model() { //加载模板 vector<Mat> myTemplate; for (int i = 0; i < 10; i++) { char name[64]; sprintf_s(name, "D:/Project/Opencv-Project/num_text/%d.png", i); Mat temp = imread(name, 0); myTemplate.push_back(temp); } return myTemplate; } void deal(Mat &src, Mat &dst, int m) { threshold(dst, dst, 100, 255, THRESH_BINARY_INV); Mat dest; resize(src, dest, Size(src.cols *1, src.rows *1), 0, 0, INTER_LINEAR); result[m].data = compare(dest, dst); result[m].num = m; } void select_result(Mat & src, int num) { for (int i = 0; i < 10; i++) { deal(src, imread_model()[i], i); } sort(result, result + 10); if (result[9].data>0.7) { cout << "第" << num << "个数字为 " << result[9].num << endl; cout << "识别精度为 " << result[9].data << endl << endl; } }