Opencv learning notes - remapping

Remapping is the process of placing pixels at a certain position in one image to a specified position in another image .

The position of each pixel is expressed by remapping (x, y)：   g(x, y) = f( h(x, y) )

g()        Is the target image ,

f( )        Is the source image ,

h(x, y)   It works on (x,y) Mapping method function .

for example ：

If there is an image A, Remapping according to this condition :  h(x, y) = (A.cols - x, y ), The image will follow x The axis direction is reversed .

remap( ) function According to the specified mapping form , Put source image Remapping geometric transformation , The formula is as follows ：

 

void remap(InputArray src,  OutputArray dst,  InputArray map1,  
           InputArray map2, int interpolation, 
           int borderMode = BORDER_CONSTANT, 
		   const Scalar& borderValue = Scalar())

Parameters 1, The source image , fill Mat Class object , And it needs to be a single channel 8 Bit or floating point image .

Parameters 2,OutputArray Type of dst, Output image , Need to be the same size and type as the source image .

Parameters 3,InputArray Type of map1, There are two possible representation objects .

              Indication point （x,y） The first mapping of .

              Express CV_16SC2 , CV_32FC1 or CV_32FC2 Type of X value .

Parameters 4,InputArray Type of map2, There are two possible representation objects , and map2 It's based on map1 To make sure

              Represents that kind of object .

                 if map1 Indication point （x,y） when . This parameter does not represent any value .

                 Express CV_16UC1 , CV_32FC1 Type of Y value （ Second value ）.

Parameters 5, Interpolation method , Previous resize() There's something about... In the function , Be careful ,resize( ) Function

                INTER_AREA Interpolation is not supported here , The optional interpolation methods are as follows ：

                INTER_NEAREST     -- Nearest neighbor interpolation

                INTER_LINEAR         -- Bilinear interpolation （ The default value is ）

                INTER_CUBIC           -- Bicubic spline interpolation （ More than 4×4 Bicubic interpolation in pixel neighborhood ）

                INTER_LANCZOS4    -- Lanczos interpolation （ More than 8×8 Pixel neighborhood Lanczos interpolation ）

Parameters 6, The boundary model , Have default values BORDER_CONSTANT, Represents... In the target image “ outliers （outliers）”

              The pixel value of is not modified by this function .

Parameters 7,const Scalar& Type of borderValue, The value used when there is a constant boundary , The default value is Scalar( ),

              That is, the default value is 0.

According to the effect ：

                   

 

Code ：

// Variable definitions 
Mat srcImage, dstImage;
Mat map_x, map_y;
// Load the original drawing 
srcImage = imread("E:\\img\\logo6.png", 1);
imshow(" The original picture srcImage", srcImage);
// Create the same rendering as the original ,x Remapping graph ,y Remapping graph 
dstImage.create(srcImage.size(), srcImage.type());
map_x.create(srcImage.size(), CV_32FC1);
map_y.create(srcImage.size(), CV_32FC1);
// Double loop , Go through every pixel , change map_x & map_y Value 
for (int j = 0; j < srcImage.rows; j++) {
	for (int i = 0; i < srcImage.cols; i++) {
		// change map_x & map_y Value . 
		map_x.at<float>(j, i) = static_cast<float>(i);
		map_y.at<float>(j, i) = static_cast<float>(srcImage.rows - j);
	}
}
// Remap 
remap(srcImage, dstImage, map_x, map_y, INTER_NEAREST, BORDER_CONSTANT, Scalar(0, 0, 0));
imshow(" Mapped image dstImage", dstImage);

Sure x Flip 、y Flip 、x y Flip And zoom . Yes map_x perhaps map_y Make changes .

map_x.at<float>(j, i) = static_cast<float>(i);
map_y.at<float>(j, i) = static_cast<float>(srcImage.rows - j);