Photoseparation system with machine vision for cleaning grain from hard-to-separate impurities

The study focuses on designing of a photo-separation system with machine vision to clean grain from hard-to-separate impurities (Tartary buckwheat). The main units of the system have been validated – a video camera, a feed hopper with a vibrator, an inclined chute, a software module, and a rejection unit. A database of wheat and Tartary buckwheat particles was developed using an industrial video camera. Using Python and the OpenCV library for frame capture, a software module was developed for recognizing grain particles and hard-to-separate impurities moving along inclined chute.