Hello!

I implemented a system where a lot of images are inputed for processing. But the problem is that the server can't handle that amount of images with a good performance, which became a really big problem for me.

The image attached is a draft of the architecture of the images. The relevant elements, for identification, are: the two QRCodes and the "box with info". For rotation and right positioning of the image, the important elements are the drawings (MARKERS) above the line.

My workaround is, apply the wavelet, catch the first quadrant and apply the processing of recognition and positioning, inverse the wavelet transform, catch the inverse image and apply the other processing in the image. I've tried a lot of different families of the wavelet transforms, but I'm losing a lot of information of the QRCODE and the markes when I caught the first quadrant.

If somebody has any ideas or tips that could help me in this problem, like different solutions or improvements that could be made in my solution, please help me.

Thank you very much!