一种SAR图像机动目标检测与分析方法

A Mobile Target's Detection and Analysis Method from SAR Image

从SAR(合成孔径雷达)图像中检测和分析目标是进行SAR自动目标识别的关键步骤,提出了一种SAR图像中地面机动目标检测与分析的方法,该方法在对图像进行预处理后首先利用背景杂波强度分布为指数分布假设的恒虚警率算法以及形态学运算对原始的SAR场景数据进行快速检测获得感兴趣的目标区域,然后提取目标区域8个特征构成特征矢量以详细描述目标。实验结果表明,该方法计算速度快,能够从获得的目标区域得到大量有用的信息,而且该方法具有一定的通用性。

The Mobile target＇ s detection and analysis from SAR（ synthetic aperture radar） image is a key step of SAR ATR（ automatic target recognition）. An approach of the mobile target＇ s detection and analysis from SAR image is proposed in this paper. After pre-processing of SAR image, Constant false alarm ratio （CFAR） algorithm based on the hypothesis that the intensity distribution of the clutter is exponential and morphological operation are applied to original SAR scene to get region-of-interest（ROIs） , the CFAR detector is a rapid algorithm for detecting targets in clutter and the morphological operation can eliminate false alarms. Then the feature vector consist of 8 features for each target area is extracted to describe target in detail. Features are length, width, area, the position of centroid, azimuth, fractal dimension, rotational inertia and Mean bright. The experimental results show that this algorithm is fast in computing speed, and can get lots of useful information from target areas. Moreover, this algorithm has definite currency.