System for Supermarket Analysis

The project is performed as a demonstration before the work on a full version begins.
The task is to determine the proportion of empty and partially empty spots in the supermarket. The recording is made with a camera that is recorded against the stallage. It is sent to the RetinaNet detector. The result of the detox is processed by an algorithm that determines the percentage ratio of filled/particularly filled/pulled columns for each columns, and for all columns on the photo. The result is in the BD.

Interesting records for each camera are pre-indicated by the user manually, for which a special utility was made (OpenCV2 + PyQT5)