关键词: 图像拼接, Wallis变换, 均值统计, 颜色空间转换, 色彩均衡

Abstract: Automatic color equalization is a very important technology for image processing. The paper proposed a new color and brightness equalization algorithm based on image pixel-mean statistical after the analysis of the problems in images mosaic, and current image color equalization algorithms which were commonly used in image mosaic were also discussed. Firstly, overlaps were extracted from two adjacent camera simultaneous frame images, and channels (RGB) were separated latterly. Then one camera image was used as a reference image, the other one as the target image, and the color channels pixel-mean were counted before the whole target image was corrected. Finally, color space was conversed (RGB to HSV) for both revised image and reference image (whole image), and their brightness channel (V channel) pixel-mean difference was calculated to correct target images brightness again. The results show that the algorithm can correct the adjacent camera image brightness and color difference effectively, and makes a good improvement for image mosaic at later period.

Key words: image mosaic, Wallis transform, mean statistics, color space transformation, color equalization