The solution was developed in Python using the TensorFlow and Keras libraries. It was trained and tested on huge number of photographs of different types of scrap metal of some industrial enterprises, moreover, some of them are visually difficult to distinguish.

Product Features:

• Highly accuracy of image classification of most types
• The applicability of the developed combinations of methods for classifying various types of images, including textural
• Implementation of the developed solution that defines the types of scrap metal from a video stream or from a photo allows to minimize the risks of false classification and the corresponding financial costs