A visual inspection system for quality control of optical lenses

Abstract

This paper proposes a quality inspection system for optical lenses using computer vision techniques. The system is able to inspect LED (Light-Emitting Diode) lenses visually and to validate their quality level automatically based on the defect severity. The optical inspection system applies the block discrete cosine transform (BDCT), Hotelling statistic, and grey clustering technique to detect visual defects of LED lenses. A spatial domain image with equal sized blocks is converted to DCT (Discrete Cosine Transform) domain and some representative energy features of each DCT block are extracted. These energy features of each block are integrated by the statistic and the suspected defect blocks can be determined by the multivariate statistical method. Then, the grey clustering algorithm based on the block grey relational grades is conducted to further confirm the block locations of real defects. Finally, a simple segmentation method is applied to set a threshold for distinguishing between defective areas and uniform regions. Experimental results show the defect detection rate of the proposed method is 94.64% better than those of traditional spatial and frequency domain techniques. Key words: Industrial engineering, quality inspection, optical lens, visual defect, computer vision system.