Click here to Skip to main content
15,887,027 members
Please Sign up or sign in to vote.
4.00/5 (1 vote)
Dear all,

This is for the my new project so i need suggestion about how to start.

I want to detect random shape defects and color changes on the fax paper. For example there is a randomly shaped torn area on the paper. I want to find this region while the paper is moving or there is a certain color change on the apper i want to find this.

My first idea is to apply edge detection like canny and then finding the contours. But the question in my head is how to find randomly shaped defects coming from production of the paper. There are a lot of example for square, triangle and circle but i did not see any for random shapes.

For the color detection, i am planning to find color of the pixels of the paper before starting the paper movement then compare color change in every pixels. Is it good idea? Because random cameras scan an area not a line. Do i need line scan camera instead of regular camera. I am not sure about how good lighting i need. I think it should be really stable wihch requires closed area.
Do you have any other suggestion?

What I have tried:

I have not tried anything yet due to wthink about the porject. Any suggestion i welcome.
Posted
Comments
Jo_vb.net 29-Nov-23 13:55pm    
This seems to become a complex [AI?] project - not sure if someone here can help you.

I would try to find a helpful sample project for shape detection here:
https://github.com/topics/shape-detection
OzzieOzburn 2-Dec-23 14:47pm    
AI is capable but not needed i guess. Anyway, thanks for suggestion.
[no name] 30-Nov-23 21:26pm    
A "random shape" needs to be defined in terms of color contrasts or a "path" of some sort if you want to get anywhere. I use color in maps to detect (at the pixel level) roads, greenery, stone buiildings, wood buildings, rivers and streams, elevations / contours. Just like they did 300+ years ago.

I "sampled" all the pixels to get a "range" for each property. At the same time by discarding any pixels not "in range", I "cleaned" the image ... or changed / enhanced colors.
OzzieOzburn 2-Dec-23 14:46pm    
Thanks for the suggestion, certainly i will consider it.
Dave Rave 9-Dec-23 10:27am    
Try with OpenCV > https://pyimagesearch.com/2016/02/08/opencv-shape-detection/

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)



CodeProject, 20 Bay Street, 11th Floor Toronto, Ontario, Canada M5J 2N8 +1 (416) 849-8900