TOWARD FINDING METHODS FOR IDENTIFYING INTERNAL AND EXTERNAL DEFECTS IN POTATO TUBERS

Abstract

The article is devoted to the relevance of developing a means of universal automatic sorting of potato tubers by external and internal defects, in particular, its main element - an optoelectronic system. Studies have shown that for sorting potato tubers it is important not only to determine their size, but also the presence of defects. The existing machines do not implement the above functions, which requires additional labor, reduces productivity and quality. The classification of quality recognition systems is given and the technological scheme of the optoelectronic system consisting of an optical module and a video camera, a control unit and a conveyor is justified. In the course of research, the two-flow type of potato flow with forced rotation and the design and technological scheme of the machine are justified. The software is developed on the basis of segmentation and tracking algorithms, as well as training of a neural network that recognizes defects of potato tubers (external and internal).