Opencv learning notes - regions of interest (ROI) and image blending

Catalog

Region of interest （ROI）

Image blending

---

Region of interest （ROI）

Select an area in the image .

Definition ROI Area There are two forms of methods ：

Method 1 ：

Use Rect, Specify the coordinates of the upper left corner and the length and width of the rectangle

Mat roi = image(Rect(100, 100, 250, 250));

image Is the loaded image .

Method 2 ：

Specify the... Of the region of interest Range of rows and columns

Mat roi = image(Range(20, 150), Range(30, 300));

Range： From the actual index to the end of the index A continuous sequence

 

According to the effect ：

  Code ：

#include <iostream>
#include "opencv2/opencv.hpp"
using namespace cv;
int main()
{
	cv::Mat image = imread("E:\\roi_test.png");
	cv::imshow(" Source graph ", image);

	cv::Mat roi = image(Range(30, 280), Range(60, 300));
	cv::imshow("roi", roi);

	cv::waitKey(0);
	return 0;
}

Image blending

First specify ROI, In use addweighted For the specified ROI The image of the region is mixed .

Linear blending is a typical binary pixel operation , The theoretical formula is ：

 g(x) = (1 - a)F1(x) + aF2(x)

      a representative alpha Value （0.0~1.0）, For two images （F1(x) and F2(x)） Or two videos （F1(x) and F2(x)） To produce the effect of overlapping pictures in time , A superimposed effect of switching from the front page to the back page .

Realization way ：addWeighted function ： Calculate two arrays （ Image array ） Weighted sum of

void addWeighted(InputArray src1, 
                 double alpha,
                 InputArray src2,
                 double beta, 
                 double gamma, 
                 OutputArray dst, 
                 int dtype = -1);

Parameters 1,InputArray Type of src1, Represents the first array to be weighted ;

Parameters 2,double Type of alpha, Express src1 The weight of ;

Parameters 3,InputArray Type of src2, Represents the second array , It needs and src1 Have Same size and number of channels ;

Parameters 4,double Type of beta, Express src2 The weight value of ;

Parameters 5,double Type of gamma, A scalar value added to the sum of the weights .

Parameters 6,OutputArray Type of dst, Output array , It has the same size and channel as the two input arrays                   Count ;

Parameters 7,int Type of dtype, Optional depth of output array , Have default values -1. When two input arrays have the same depth                 When the degree of , This parameter is set to -1（ The default value is ), That is equivalent to src1.depth().

Two arrays (src1 and src2 ) Weighted sum of :    Yes addWeighted Parameters in beta Position as 1 - alpha   

gamma Position as 0

dst = src1[ i ] * alpha  +  src2[ i ] * beta  +  gamma;

i： Index values of multidimensional array elements , In the case of multi-channel arrays , Each channel needs to be handled independently . When the depth of the input array is CV_32S when , This function does not apply .

According to the effect ：

 

Code ：

Mat srcImg1;
Mat srcImg2;
Mat srcImg3;
	
srcImg1 = imread("E:\\img\\logo3.png");
srcImg2 = imread("E:\\img\\logo4.png");

imshow(" Source graph srcImg1", srcImg1);
imshow(" Source graph srcImg2", srcImg2);

addWeighted(srcImg1, 0.5, srcImg2, 0.5, 0.0, srcImg3);
imshow(" Mixed srcImg3", srcImg3);

You can also specify roi after , Then mix a certain area in the picture .