How to recognize a defect in an image?

Question

Images

There is a task of recognizing a defect in a weld on an automatic line. Visually, it differs from a weld without a defect in the presence of black. The image is obtained by a video camera, it is required to process this image in real time and issue a message when a defect is detected in the image. Can anyone suggest what programming language can be used to solve this problem? Or maybe he knows and can suggest companies or people from whom you can order this work?
I studied the literature on image processing, I can imagine how the algorithm works, but I still don’t really understand how it can be done

Answer 1

Python/C++ and OpenCV. With a suitable camera, you can probably simply determine by the histogram.

Answer 2

The task is specialized and requires writing a separate algorithm, we can say that the solution is tailored to the task. It is solved, as written above using OpenCV for C ++, you can use EmguCV (wrapper for .NET for OpenCV)

Answer 3

If everything is as described, you can do without any CV :-)
Write in any language you know, get an image from the camera, build a histogram of brightness in rows/columns or even the entire image, and study the distribution... (this is not only possible find defects, but also evaluate the uniformity of the seam)
... and if we simplify even more, then we replace the camera with a photodiode, and select the backlight so that we get contrast for craters \ slag \ scale, etc.
PS. Write in a personal if anything, I will do / advise ...