scratch or defect (?) detection

10 次查看（过去 30 天）

显示更早的评论

emre parlak 2018-11-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/430316-scratch-or-defect-detection

评论： Image Analyst 2018-11-16

hey guys.

i have this photos that i need to find the scratches or defects on them. as an example:

as you can see (i hope you can see) there is a scratch on the upper side of the white band there.

i don't know that much about image processing, but when i surf on this forum, i came to this point:

i used edge detection to find objects on the image, my goal was to determine the objects and eleminate them, hence acquiring an image without any obsticles and only have scratch on it. and then using line detection, or using gradients, find that scratch.

but i couldn't manage to fill in the lines on that BW image. am i doing this hard way? is there a source you guys can forward me to?

any help would be appreciated.

thanks

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

回答（1 个）

Image Analyst 2018-11-16

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/430316-scratch-or-defect-detection#answer_347382

You need to contact a professional to get your image capture situation fixed. Those are horrible images. If you want to find scratches in there, you need to fix your lighting problems, focus problems, resolution problems, and geometry problems. Once that's done you can start doing the image analysis. But with your current setup, there's little hope to find faint scratches anywhere.

5 个评论
显示 3更早的评论隐藏 3更早的评论

zaydon latif 2018-11-16

Image Analyst sir , Do you have an email to talk to you? I want to help you .

Image Analyst 2018-11-16

No, sorry. I don't need help. Like I said, it's you who should seek the help of a professional computer vision expert to help you design an optical system to capture good images. What you have now is junk. I do this all the time (I have a Ph.D. in optics) but you should probably get someone at an industrial machine vision or scientific imaging company in your own city to help you.

请先登录，再进行评论。

请先登录，再回答此问题。