Research on Multi-Channel Metal Bottle Cap Defect Detection System Based on Machine Vision

This paper presents an advanced defect detection system for bottle caps, addressing challenges in quality control within the production process. Traditional methods struggle with irregular cap shapes, inconsistent lighting, and various defects, such as dark spots, color mixing, scratches, and missing gaskets. The proposed system integrates Low-Angle and Large-Angle Differential (LALDA) lighting with Multi-Channel Segmentation (MCS) to enhance detection accuracy and efficiency. The LALDA system utilizes a low-angle light source to uniformly illuminate the cap's sidewall and inner circular region, mitigating issues of uneven lighting and background interference. The MCS algorithm, based on HSV color space, automatically segments the cap into distinct sub-regions, enabling precise detection of defects in each area. Specifically, the system utilizes adaptive thresholding, binary operations, and contrast enhancement to detect dark spots, color mixing, scratches, and missing gaskets. Experimental results demonstrate the system's effectiveness, with improved segmentation accuracy and robustness to variations in product positioning and lighting fluctuations. This approach offers a promising solution for automated, high-precision defect detection in bottle cap manufacturing, significantly contributing to enhanced product quality and production efficiency.