| 基于多特征提取的遥感图像机场目标自动检测 |
| |
| 引用本文： | 王彪,姜志国,赵丹培. 基于多特征提取的遥感图像机场目标自动检测[J]. 中国体视学与图像分析, 2009, 0(2): 120-124 |
| |
| 作者姓名： | 王彪 姜志国 赵丹培 |
| |
| 作者单位： | 北京航空航天大学宇航学院图像处理中心,北京100191 |
| |
| 基金项目： | 基金项目：国家自然基金项目（69776793） |
| |
| 摘 要： | 传统的遥感图像机场跑道自动目标检测由于仅提取灰度特征常产生过分割现象,本文采用灰度特征和纹理特征相结合的方法进一步提高跑道的检测精度。利用阈值对遥感图像进行初始分割,以定位感兴趣区域（ROI）,再利用EM算法估计ROI区域训练样本,引入马尔可夫随机场（MRF）模型,分割机场跑道。实验表明MRF可以很好地描述空间连续性,可以达到精确检测机场跑道的目的。
|
| 关 键 词： | 灰度特征 纹理特征 阈值分割 EM算法 马尔科夫随机场（MRF） |
Automatic target detection of airfield runway in remote sensing image by multi-feature extraction |
| |
| WANG Biao,JIANG Zhiguo,ZHAO Danpei. Automatic target detection of airfield runway in remote sensing image by multi-feature extraction[J]. Chinese Journal of Stereology and Image Analysis, 2009, 0(2): 120-124 |
| |
| Authors: | WANG Biao JIANG Zhiguo ZHAO Danpei |
| |
| Affiliation: | (Image Processing Center, School of Astronautics, Beijing University of Aeronautics ＆ Astronautics, Beijing 100191, China) |
| |
| Abstract: | Traditional automatic methods for airport runway target detection in remote sensing image use only gray feature. This paper presents a method which combine texture feature with gray feature to improve the precision of runway target detection. To extract the region of interest （ ROI）, the method uses thresh- old to get the initial segmentation and then estimate training samples of ROI by EM algorithm. Followed by this, the Markov Random Fields （MRF） model is employed to make the segmentation in the ROI to detect the runway target. Experiment results show that MRF has good performance in describing spatial continui- ty. It can well describe the airport runway and get accurate detection result. |
| |
| Keywords: | gray feature texture feature threshold segmentation EM Algorithm Markov random fields |
| 本文献已被 维普 等数据库收录！ |
