图像变化检测照度变化和小偏移的问题

时间:2018-09-04 17:22:36

标签: python c++ image-processing

“不受不同照明,偏移和方向影响的图像变化检测”的最佳解决方案是什么?

Python或c ++很好

附加的示例图片:

original and second image

image

image_

1 个答案:

答案 0 :(得分:1)

您可以将图像对齐与功能匹配配合使用,例如:

//Load 2 Slices with Body Region  
 img1 = LoadImageBody("Slice01");   
 img2 = LoadImageBody("Slice02");  
 //Calculate Keypoints&Descriptors  
 KeyPt1 = SURF(img1);  
 KeyPt2 = SURF(img2); 
 //Find Corresponding Point Pairs   
 Pairs = FindPairs(KeyPt1, KeyPt2); 
 //Find Homography Matrix   
 H = FindHomography(Pairs); 
 //Calculate Image Transformation  
 imgWarp =WarpPerspective(img2, H); 
 //Save Transformation Result   
 SaveImg("Warp.png", imgWarp); 
 //Calculate Two Images Difference 
 imgDiff = AbsDiff(img1, imgWrp); 
 //Save Difference Result
 SaveImg("Diff.png", imgDiff); 

但是我没有解决不同的照度和颜色变化的方法。.我也在寻找..