Normalize light
Using a webcam. With optimized light (by trial and error) for OpenCV it recognize what it should, in this case playing cards but could be any object. As soon as the light changes in any way, the recognition fails. Is there any way to have OpenCV to automatically adjust for different light situations?
Thats probably because you trained a system to only recognize well illluminated objects. Without a deeper insight in your processing pipeline we cannot help you in finding decent solutions.
Thanks for your answer. I understand your point. I trained the system with a well-lit photo of all playing cards. Of course it fails when light is less or different.
" I trained the system" -- what does it mean, exactly ?
Using this: https://github.com/vascozzz/Augmented...