Tuesday, 15 July 2014

c++ - OpenCV: Denoising image / video frame -


I want to denoise a video using OpenCV and C ++ I found on the OpenCV Doctor site:

  fastNlMeansDenoising (contourImage, contourImage2);  

Each time a new frame is loading, my program should represent the current frame and it should be written in compaired IMS2.

But if I run the code, 0 returns and goes out what I am doing wrong or is there an alternative way to denoise an image? (It should be fast, because I am processing a video)

When you use c ++ If you are using, you are not trying to provide full arguments like this.

  cv :: fastNlMeansDenoisingColored (contourImage, contourImage2, 10, 10,7, 21); // The basic function to use is Cv :: fastNlMeansDenoising (src [, dst [, h [, templateWindowSize [, searchWindowSize]]]]] DST  

parameter:
src - Input 8-bit 1-channel, 2-channel or 3-channel image.
DST - Type the output as the image and source with the same size.
Template WindowSize - The size in pixels of the template patch whose weight is used to weigh should be the recommended value of 7 pixels.
searchWindowSize - To calculate the weighted average for a given pixel, the size of pixels in the window used to be peculiar should be affecting the linear effect: more.
searchWindowsSize - maximum denoising time. Recommended value is 21 pixels. Control the filter strength of the h - parameters; Big H value completely removes the noise, but also removes the image details, the small H value keeps the details safe, but Some noise also preserves


No comments:

Post a Comment