Opencv learning warp Perspective

One 、 Get perspective transform matrix function GetPerspectiveTransform

CV_EXPORTS Mat getPerspectiveTransform(const Point2f src[], const Point2f dst[], int solveMethod = DECOMP_LU);

• Parameters src Source image quadrilateral vertex coordinates .

• Parameters dst Quadrilateral vertex coordinates corresponding to the target image .

• Parameters solveMethod Pass to cv::solve(#DecompTypes) The calculation method of , The default is DECOMP_LU

notes ： At this point, we get a matrix （ Perspective transformation matrix ） 

Two 、 Perspective transformation function warpPerspective

• Parameters ：src The input image .

• Parameters ： dst Output image , have Size dsize And it has the same type as the source image .

• Parameters ： M 3*3 The transformation matrix of .

• Parameters ： dsize The size of the output image .

 

float w = 250, h = 350;

int main() {
	string path = "C:\\Users\\Rong\\Desktop\\Resources\\cards.jpg";
	Mat Img = imread(path);

    // Two Point2f Array of ！
    // Pay attention to the src and dst In the same order ！
	Point2f src[4] = { {743,384},{1021,435},{965,783},{648,710} };
	Point2f dst[4] = { {0,0},{w,0},{w,h},{0,h} };

	Mat matrix,ImgWarp;
	matrix = getPerspectiveTransform(src, dst);
	warpPerspective(Img, ImgWarp, matrix, Size(w, h));

	for (int i = 0; i < 4; i++) {
		circle(Img, src[i], 10, Scalar(0, 0, 255), FILLED);
	}

	imshow("Image", Img);
	imshow("Image Warp", ImgWarp);
	waitKey(0);


}