CAJ | 학술논문

针对彩色多聚焦图像融合存在色彩分布不均、斑效应问题，提出一种图像空间信息和颜色信息分离研究的改进算法。首先对源图像进行 HSV 变换并分离各分量，然后表示空间信息的 V 分量进行小波分解，最高分解层高频系数选用基于邻域的对比度取大融合规则，而其余分解系数选取基于邻域梯度的加权融合规则。其次重构得到融合图像 V 分量，表示颜色信息的 H、S 分量采用源图像与融合图像间 V 分量欧氏距离取小规则，最后进行 HSV 逆变换得到融合图像。经 MATLAB 仿真，该算法可有效缓解融合图像色彩分布不均及斑效应现象。
침대채색다취초도상융합존재색채분포불균、반효응문제，제출일충도상공간신식화안색신식분리연구적개진산법。수선대원도상진행 HSV 변환병분리각분량，연후표시공간신식적 V 분량진행소파분해，최고분해층고빈계수선용기우린역적대비도취대융합규칙，이기여분해계수선취기우린역제도적가권융합규칙。기차중구득도융합도상 V 분량，표시안색신식적 H、S 분량채용원도상여융합도상간 V 분량구씨거리취소규칙，최후진행 HSV 역변환득도융합도상。경 MATLAB 방진，해산법가유효완해융합도상색채분포불균급반효응현상。
Multi-focus color image fusion has the problems of color maldistribution and spot effect, an im-proved fusion algorithm is put forward, which separates spatial and color information of image. Firstly, the source images makes HSV transform and separates their components, then the V component representing space information decomposes wavelet, and the high frequency coefficients of the highest layer are fused a-dopting larger neighborhood contrast, while the other coefficients are selected based on neighborhood gradi-ent weight. Then the V component is reconstructed, the H and S components, which reflect image’s color information, are fused by smaller Euclidean distance between source and fusion images’V component. Fi-nally fusion image is achieved through HSV inverse transformation. Simulations in MATLAB indicate that the algorithm could effectively ease fusion image color maldistribution and spot effect phenomenon.