Software system for detecting brick defects based on computer vision methods

Abstract

For industrial enterprises, an important area of activity is the use of non-destructive testing in assessing product quality. One of the most developed approaches in practice are optical methods (visible or infrared spectral regions) and associated algorithmic approaches to computer vision. Today a large number of industries use a manual method for detecting defects, which is labor-intensive, slow and subjective. The article presents a description of an automated optical control system to detect imperfections in bricks on a conveyor belt during their production. The optical control is based on RGB images from digital cameras. The processing algorithm is shown and the results obtained are presented