非下采样轮廓波变换的前视声呐图像融合  

Forward-looking sonar image fusion based on non-subsampled contourlet transform

作　　者：张健[1,2] 卞红雨 ZHANG Jian BIAN Hongyu(College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin 150001, China Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China)

机构地区：[1]哈尔滨工程大学水声工程学院,黑龙江哈尔滨150001 [2]哈尔滨工程大学水声技术重点实验室,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2017年第9期1373-1379,共7页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(41376102)

摘　　要：针对前视声呐图像清晰程度不同,局部区域模糊的特点,本文提出一种基于非下采样轮廓波变换的前视声呐图像融合算法。依据图像多尺度分解的理论,对源图像进行非下采样轮廓波变换,得到一系列多尺度子带分解系数;根据图像中清晰目标反射声波能量大、对比度高特点,构建前视声呐图像融合规则,即低频子带采用Gabor能量、高频子带计算局部对比度指导融合规则,提出区域一致性校验准则抑制图像噪声,产生融合图像多尺度子带分解系数,并应用非下采样轮廓波逆变换获得融合图像。声呐图像融合对比实验证明,采用提出方法生成的融合图像在主观视觉和客观指标上均优于其他融合方法。According to the inherent properties of forward-looking sonar （FLS） images, such as non-uniform clarity and local blurred regions, a novel fusion method based on non-subsampled contourlet transform is proposed for FLS image fusion. Based on the theory of muhi-scale analysis, the source images are decomposed into multi-scale coefficients by non-subsampled contourlet transform; the fusion rules are constructed on the fact that the clear objects have high reflection energy and contrast in FLS images, therefore the low-frequency subband is fused by the Gabor energy and the high-frequency sub-bands are fused by the local contrast, respectively; the region consistency check is then implemented to suppress the influence of noise; the fusion image is obtained by inverse non-subsampled contourlet transform on the generated fused coefficients. Experiments on FLS image fusion have been conducted to demonstrate the effectiveness and the superiority of the proposed technique using both a subjective evaluation and objective metrics.

关 键 词：图像融合 多尺度分析 非下采样轮廓波变换 前视声呐图像 融合规则 Gabor能量 局部对比度 区域一致性校验 

分 类 号：TN911.73[电子电信—通信与信息系统]