1 mstar图像灰度分布统计分析 - 中国图象图形学报.doc

中图法分类号：TP391.1 文献标识码：A 文章编号：1006-8961(2013) 论文引用格式：倪维平，严卫东，吴俊政，郑刚，芦颖. MSTAR图像的矩特征分析与多阈值分割[J] MSTAR图像的矩特征分析与多阈值分割 倪维平1,2，严卫东1,2，吴俊政2，郑刚2，芦颖2 （1.西北核技术研究所，西安 710024，2.西安电子科技大学电子工程学院MSTAR image and multiple thresholds based segmentation Ni Weiping, Yan Weidong, Wu Junzheng, Zheng Gang, Lu Ying (1.Northwest Institute of Nuclear Technology, Xi’an, 710024, China，2.School of Electronic Engineering, Xidian University, Xian 710071, China) Abstract: Image segmentation is a fundamental step for the SAR image based automatic target recognition (ATR). A method based on moment feature and multiple thresholds is proposed for MSTAR image segmentation. Firstly, according to a comprehensive research of the statistics of MSTAR images, mathematical description models for regions of target, shadow, and background are constructed respectively. Then the moment feature is defined, followed with the analysis of its basic properties. By transforming the image into moment feature space, the difference between target region and the other two types of regions is significantly enhanced. Finally, a strategy with multiple thresholds is constructed to fulfil the segmentation. The experiment results with MSTAR dataset indicate that the algorithm presented here has advantages not only on the noise robustness, but also on the segmentation effect, the processing efficiency over the common-used methods, such as OTSU, FCM, MRF, and CFAR. Furthermore, this new method also performs well in the segmentation of MSTAR images with various scales and multiple targets. Key words: MSTAR image; image segmentation; moment feature analysis; multiple thresholds 0 引 言 合成孔径雷达（SAR）是近年来取得快速发展的一种全天时、全天候对地观测手段[1]，因其应用领域广阔而备受关注 [2]，其中SAR图像自动目标识别（ATR）是重要领域之一。1995年，美国国防高级研究计划署（DARPA）和空军研究室（AFRL）公开发布的运动、静止目标获取与识别（MSTAR）数据集，有力推动了目标识别技术的发展。典型的SAR图像ATR系统包括目标辨识（target discrimination）、目标分类（target classification）和目标识别（target recognition）三个主要部分，其中目标辨识又称为目标检测，主要用于区分目标和背景，多采用图像分割来实现。 图像分割是根据不同的规则，将图像划分成具有不同特性（如纹理、灰度等）的区域。现