# Does anyone know a suitable algorithm to deal with such a complicated problem?

I intend to find cracks in the image. However, as you can see, the noise is very heavy. So, does anybody know how to cope with such a extreme situation? Thanks in advance.

edit retag close merge delete

1

I wonder if a hough circle function could be modified to allow for variations of arc during the line.

( 2014-03-08 19:11:21 -0500 )edit

I have the same question as yours. Suppose it can, then how to reduce the effect caused by such heavy noise?

( 2014-03-11 06:52:31 -0500 )edit

The crack is not a perfect line but it does have some sections that are less noisy. I would think you would have to start with a Gaussian blur function. Then the hough functions start by measuring local line angles and then tries to stitch them together into a longer line. The noisy bits won't pass this second part of the process.

( 2014-03-12 03:13:35 -0500 )edit

Sort by ยป oldest newest most voted

This image is color, so the other method is try to convert the other color space and find the easy inspection image. My suggestion try:

boxfilter(src,temp1,kernel......)  //try bigger core. or you can use frequency filter
temp2=cv::substract(src,temp1)


or

   boxfilter(src,temp12,kernel1......)   // the kernel size kernel1<kernel2
boxfilter(src,temp11,kernel2......)  //try bigger core. or you can use frequency filter
temp2=cv::substract(temp11,temp12)


then try to

threshold (temp2,temp3......)  //or the orher threshold


finally try to use blob to find the longest line or characteristic which is near horizontal.

best regard

more

It doesn't work well. Not much noise can be removed.

( 2014-03-11 08:37:08 -0500 )edit

the kernek size should be >11 you can change the other filter and try again and may be you must use area of blob to filter some noise.

( 2014-03-11 08:44:06 -0500 )edit

Official site

GitHub

Wiki

Documentation

## Stats

Asked: 2014-02-28 22:21:34 -0500

Seen: 269 times

Last updated: Mar 02 '14