Vision-Based Angle Detection for Industrial Quality Inspection

Accurate geometric inspection is a fundamental requirement in modern industrial quality control, especially in robotic assembly lines where precision dictates product integrity. This paper introduces a vision-based system designed to detect and measure angles between structural features using a single camera. The proposed approach combines edge detection (Canny), Hough transform, and vector analysis to identify lines and compute angles in real time. The system allows interactive adjustment of detected points, enabling fine-tuning for high-accuracy measurements during setup or validation stages. By providing non-contact, automated angle verification, this solution enhances inspection efficiency in manufacturing processes such as metal sheet bending, welded joint evaluation, and mechanical part alignment. Experimental evaluations conducted in factory-like environments demonstrate a measurement accuracy within ± 1.092°, with resilience to moderate noise and lighting variability. This system provides a cost-effective alternative to laser or tactile devices, making it highly adaptable for robotic inspection tasks in Industry 4.0 contexts.