基于奇异值分解与同态滤波的粮虫图像增强  

Pest Image Enhancement Based on Singular Value Decomposition and Homomorphic Filter

作　　者：牟怿[1] 周龙[2] 

机构地区：[1]安徽科技学院工学院,凤阳233100 [2]武汉工业学院电气信息工程系,武汉430023

出　　处：《中国粮油学报》2011年第2期108-110,114,共4页Journal of the Chinese Cereals and Oils Association

基　　金：湖北省教育厅青年基金(Q200618002)

摘　　要：储粮害虫的在线检测是研究的热门领域。由于图像处理技术逐渐成熟,基于图像处理的储粮害虫检测系统得到了广泛的关注。图像增强是特征提取前重要的预处理过程,因此在基于图像处理的检测系统中是最重要的步骤。介绍了同态滤波图像增强方法和矩阵奇异值分解的基本理论,在此基础上提出了基于奇异值分解和同态滤波的粮虫图像增强算法。为了验证方法的有效性在Matlab平台上做了仿真试验,使用本方法增强粮虫图像并和其他方法做了对比,结果表明该方法对于粮虫图像增强是有效且可行的。Stored- grain pest online detection is an always discussed field, the pest detection system based on image processing is wildly concerned along with the image processing technology becoming mature gradually. Image enhancement is a crucial pre - processing before feature extraction, so it is the most important process of grain pest on- line detection system using image processing. The fundamental of singular value decomposition and homomorphic filter are introduced. A novel image enhancement method based on the fundamental was proposed. To verify the effectiveness, experiments on Matlab platform were carried out, i. e. , using the proposed method to enhance image and to contrast it with other algorithms. The experiments manifested that the proposed approach was efficient and feasible.

关 键 词：图像增强 奇异值分解 粮虫图像 同态滤波 

分 类 号：TN911.73[电子电信—通信与信息系统]