为了获取完整且大面积的印刷电路板(PCB)光电图像,在淡入淡出的自然缝合算法的基础上,提出了一种基于硬件的图像自动拼接新算法,并分析了其基本原理。选择用中值滤波滤除图像中的噪声。针对获取后的PCB光电图像整体偏暗的特点和其直方图中出现灰度冗余的情况,提出了去灰度冗余与灰度变换结合起来的图像增强新算法,并分析了其基本原理。用图像自动拼接算法得到了完整且大面积的PCB光电图像实验结果;用完整图像中的局部小图像从主观与客观两方面对几种图像增强算法的增强效果进行了比较,本图像增强算法的增强效果较好。 In order to get a complete and large area printed circuit board (PCB) optoelectronic image, a new algorithm of automatic image stitching based on hardware is proposed based on fade-in and stitching algorithms. The basic principle of the method is analyzed. Choose to use median filtering to filter out noise in the image. Aiming at the characteristics of darker overall PCB image and the appearance of gray-level redundancy in the histogram, a new image enhancement algorithm combining gray-scale redundancy with gray-scale transformation is proposed and its basic principle is analyzed . The experimental results of complete and large-area PCB photoelectrical images are obtained by the automatic image stitching algorithm. Compared with the enhancement effect of several image enhancement algorithms, the local small images in the complete image are compared in both subjective and objective ways. Enhance the better.