摘要
该文首先介绍SAR图像分割的概念,分析了其地物数据的统计特性,在此基础上利用多种模型构成的统计模型组来匹配大幅SAR图像中各类地物的直方图分布,给出了衡量实际地物直方图和假设已知模型匹配程度的检验统计量,以此来选取最优的统计模型组;并提出了基于统计模型组的Markov随机场的SAR图像分割算法,利用Radarsat的实测数据验证了算法的有效性,给出了性能评估结果,并与其它分割方法做了比较。
This paper firstly introduces the concept of SAR image segmentation, and analyses the statistical properties of local terrain data in this image. The multi-model notion which includes some different models fitting the different terrains is introduced to describe accurately the local region in the large scene. The statistics for testing good-of-fit between the histogram of the sample data and the known distribution under the model assumptions are proposed, which are used to select the optimum models of the local terrains. SAR image segmentation is achieved based on Markov random field with multi-model for the all kinds of region in the image. The real SAR image data from Radarsat is applied to validate the above algorithm, and the performance of segmentation is discussed and compared with the other method of segmentation.
出处
《信号处理》
CSCD
北大核心
2008年第2期272-276,共5页
Journal of Signal Processing
