基于阴影的SAR图像车辆目标三维特征提取

3-D Feature Extraction of Vehicle Target from SAR Imagery Based on Shadow Information

目标长、宽、高三维几何特征,对SAR图像解译与目标识别等具有重要意义。从SAR成像几何出发,根据SAR观测俯仰角、目标及阴影之间的几何关系,研究提出了基于两维高分辨SAR图像阴影信息的车辆目标三维几何特征提取方法。该方法既可由某一方位角的单幅SAR图像提取目标三维几何特征,也可通过任意有限(3～5)个方位角SAR图像的融合提取目标三维几何特征,而且融合还可有效提高目标三维几何特征的提取精度。通过大量MSTAR实测SAR图像数据的实验结果,验证了其有效性。

Three-dimensional geometric feature（length, width and height） of a target is very important for SAR imagery interpretation and target recognition. Beginning with the analysis of SAR imaging geometry and according to the geometric relations between the depression of SAR platform, target and shadow, a novel method of 3-D geometric feature extraction of vehicle target from 2-D high resolution SAR imagery based on shadow information is proposed in this paper. Not only the method can extract 3-D geometric feature of target from single SAR image, but also can from the fusion of a very few（just 3 to 5） SAR images with different azimuth, and the fusion can improve the precision of 3-D geometric feature extraction remarkably. The validity of the proposed method has been proven by a large number of experimental results using MSTAR measured SAR imagery.