结合灰度共生矩阵的多分辨率图像融合策略  

Fusion strategy of multi-resolution image combined with gray level co-occurrence matrix

作　　者：阳魁[1] 黄新生[1] 徐婉莹[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院自动控制系,长沙410073

出　　处：《计算机应用研究》2012年第3期1181-1184,共4页Application Research of Computers

摘　　要：多分辨率图像的融合策略是提升融合图像质量的关键。目前常用的融合策略都没有直接考虑融合图像的纹理信息,根据灰度共生矩阵(GLCM)能较好地反映纹理粗糙度和方向的特点,提出了一种结合灰度共生矩阵的多分辨率图像融合策略。首先计算多分辨率图像的低频和高频部分的灰度共生矩阵,然后将低频图像GLCM的能量和高频图像GLCM的对比度对应作为低频和高频系数选取的活性度量进行图像融合。实验表明,所提出的融合策略与目前常见的融合策略相比,在主观感官和客观评价上都能获得较好的融合效果。Fusion strategy of muhi-resolution image is the key to enhance the quality of integration image. The most commonly used strategies are not directly consider the image texture information. According to GLCM can reflect the characteristics of grain roughness and direction accurately, this paper proposed a fusion strategy of multi-resolution image which combined with GLCM. First,it calculated the GLCM of the low and high frequency part of the multi-resolution image. Then the energy of the low-frequency GLCM and the contrast of the high-frequency GLCM corresponded as activity measure for image fusion to selecte the low and high frequency coefficients. Experiments show that compared to the current common fusion strategy, the proposed method can make fusion effect better through subjective sensory and objective evaluation.

关 键 词：灰度共生矩阵 多分辨率分解 图像融合 活性度量 融合策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]