Considering the different influence of image segmentation about the factorsof the distance between target and background as well as each kind of cohesion, an improved Otsualgorithm is proposed, and its basic principle and segmentation advantages are analyzed in detail. Inorder to uate these segmentation results impersonally by using different algorithms, the quantitativecriteria of gray-level contrast and district interior uniformity are adopted to uate these segmentationresults impersonally. Finally, the different segmentation experiment contrasts of PCB photoelectric imagebetween our algorithm and other algorithms is d, the results of experiment indicate that ouralgorithm has relatively better segmentation quality-Considering the different influence of image segmentation about the factorsof the distance between target and background as well as each kind of cohesion, an improved Otsualgorithm is proposed, and its basic principle and segmentation advantages are analyzed in detail. Inorder to　uate these segmentation results impersonally by using different algorithms, the quantitativecriteria of gray-level contrast and district interior uniformity are adopted to　uate these segmentationresults impersonally. Finally, the different segmentation experiment contrasts of PCB photoelectric imagebetween our algorithm and other algorithms is　d, the results of experiment indicate that ouralgorithm has relatively better segmentation quality