基于小波变换与图像不变矩的遥感图象匹配研究  

A study of remote sensing image matching based on combination of wavelet transformation and image invariant moment

作　　者：陈宇玺[1] 韩崇昭[1] 王明军[2] 康欣[1] 

机构地区：[1]西安交通大学电信学院,陕西西安710049 [2]西安建筑科技大学,陕西西安710055

出　　处：《电波科学学报》2004年第4期444-447,共4页Chinese Journal of Radio Science

基　　金：国家 973项目资助 (2 0 0 1CB30 94 0 3)

摘　　要：小波变换具有数据压缩和检测局部突变的能力 ,而图像空间不变矩对于平移、旋转、比例缩放、灰度对比度变化具有不变性。两者优势的结合不但可以有效减少遥感图像匹配的计算量 ,而且还可以减弱对于图像采集平台拍摄方位、拍摄距离、角度气象条件等的依赖性 。A study on remote sensing image matching based on combination of wavelet transformation and image invariant moment is discussed. The wavelet transformation is capable of compressing the image and detecting its mutation. The image spatial invariant moment is well applicable to image translation, rotation, scale variation and contrast change. Combination of these two methods can effectively reduce computation cost for image matching and the dependence on image acquisitive position, direction, atmosphere condition etc. It improves the algorithm applicability.

关 键 词：小波变换 图像不变矩 遥感图象 目标识别 图像匹配 

分 类 号：TP752[自动化与计算机技术—检测技术与自动化装置]