利用OpenCV进行图像配准

项目代码：GitHub：team79/ImageRegistration 
图像配准类： 
文件： 
ImageRegistration.h 
ImageRegistration.cpp

class ImageRegistration {
    Mat src1;
    Mat src2;
    void getRegistrationMat();
    Mat transMatrix;
public:
    void init(Mat,Mat);
    Point2f getXY(int,int);
    ~ImageRegistration();
};

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10

其中src1为原图像，src2为待匹配图像。 
初始化：

Mat src1 = imread("test-1.png", 1);
Mat src2 = imread("test-2.png", 1); 
ImageRegistration llll;
llll.init(src1, src2);

• 1
• 2
• 3
• 4

调用getXY函数即可得到原图中的点在src2中的坐标

Point2f temp = llll.getXY(i, j);

• 1

实际演示： 
比如我们将src1中的（100，100）起的50*50的块涂黑，对应的src2中也在相应位置涂黑

for (int i = 100; i < 150; i++) {
    for (int j = 100; j < 150; j++) {
        Point2f temp = llll.getXY(i, j);
        int temp1 = temp.x;
        int temp2 = temp.y;
        src1.at<cv::Vec3b>(j, i)[0] = 0;
        src1.at<cv::Vec3b>(j, i)[1] = 0;
        src1.at<cv::Vec3b>(j, i)[2] = 0;

        src2.at<cv::Vec3b>(temp2, temp1)[0] = 0;
        src2.at<cv::Vec3b>(temp2, temp1)[1] = 0;
        src2.at<cv::Vec3b>(temp2, temp1)[2] = 0;
    }
}
//warpAffine(src1, src2, mat, src2.size());
cv::namedWindow("Display window1", cv::WINDOW_AUTOSIZE);
cv::imshow("Display window1", src1);
cv::namedWindow("Display window", cv::WINDOW_AUTOSIZE);
cv::imshow("Display window", src2);
waitKey(0);
src1.release();
src2.release();

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22

效果：