关键词: 自动光学检测, 布尔矩阵编码, 局部最优, 修复策略, &ldquo, 两段式&rdquo, 遗传算法

Abstract: The number of images and the track affect the detection efficiency directly in automatic optical inspection; the image stitching method is less efficient. The problem is decomposed into clustering and trajectory planning problems. A genetic algorithm based on Boolean matrix coding is developed to realize clustering. This method can not only ensure that the individual is locally optimal, but also can respond to the constraint conditions by the repair strategy. Then, the “two segment” genetic algorithm is proposed to solve the position and the trajectory of the image taking, and the optimization process of the location and trajectory of the image taking is realized. Experimental results show that the location of the image and the trajectory of the image are improved compared with the traditional algorithm, which can improve the efficiency of automatic optical detection.

Key words: automatic optical inspection, Boolean matrix coding, local optimum, restoration strategy, two stage genetic algorithm