Computer vision systems are widely implemented in automatic inspection systems. The quality management of mechanical parts in industries is vital for proper functioning of machineries. Defect detection should be done in pre-production stage ensuring quality control. Real time inspection using manual labour is inadequate, time consuming and non-consistent. Hence there is a need for a system which is built for automatic defect detection, such that it avoids human errors and is comparatively

more »

... te. The system builds a computer vision system which detects the defective objects and segregates it. This paper makes use of an overhead camera mounted at specified height over a conveyor belt, which sends recorded images to the Raspberry Pi. Pattern recognition is performed using Open CV to identify the defective objects moving over a conveyor belt. It identifies defective number of teeth in gears and surface abrasions in metal sheets and thereby helps in quality management.