A Novel Computer Vision Framework for the Automated Visual Inspection for Quality Control of Automotive Fasteners

Abstract

Object recognition, a fundamental task in computer vision, plays a pivotal role in various manufacturing industries of any domain like pharmaceutical, Automobile, Robotics, surveillance electronics, food and many more. Even though, there exists a wide requirement in object recognition, need for the same is ear marked in the field of automobile as this manufacturing has to undergo lots and lots of the processing steps. Quality of the parts needs to be ensured in each and every step for the 100% quality pure components outlet. This is ensured through the process of object detection and recognition through vision systems. Key challenges in this area may vary from part detection to feature extraction in a wider scope. However, identifying the smaller components is still there in the top of the object recognition challenges. This research addresses one of such challenges of detection and recognition of small automobile parts like fasteners and it is achieved through the cutting-edge technologies. A novel algorithm has been proposed for the effective detection and classification of fasteners even at the lower illumination. The proposed Computer vision system uses vision technologies which are currently gaining prominence, enabling precise dimension measurement and enhancing the ability to recognize fasteners in complex environments. Using this proposed methodology present in this research leads to the effective classification and dimensional analysis of fasteners, specifically bolts and nuts with an accuracy level of 90% through appropriate imaging. The objective of the proposed work is to automate the quality process check in a production line of fasteners manufacturing which enables the industry to optimize their production.