基于方位不变性的SAR图像旋转体检测

针对复杂环境中微弱旋转体目标(如地雷等)检测的难题,提出了一种基于方位散射特征和局部对比度特征融合的检测算法。首先,对旋转体目标特性进行了分析,进而利用子孔径SAR图像提取方位散射熵作为待检测特征。对全孔径SAR图像分别进行方位不变性检测和CFAR检测,并将检测结果相融合,得到最终检测结果。算法体现了利用目标先验知识辅助检测的思路,实测数据结果表明,该方法能够有效剔除原先在全孔径图像中无法剔除的杂波,有效降低检测的虚警率。

Body-of-Revolution Target Detection Algorithm Based on Aspectual Invariance in SAR Image

In complex environments, detection of weak body-of-revolution（BOR） targets, such as land- mines, is a difficult problem. A detection approach of BOR targets is proposed in this paper, which exploits the characteristics of aspectual invariance and local contrast. Using sub-aperture image, characteristic of BOR targets is analyzed and azimuth scattering entropy（ASE） is then extracted as a detection feature. The ultimate detection result is obtained by fusing results from aspectual invariance detection and CFAR detection of the full aperture image. The idea of detecting targets with priori information is showed in this paper. The results of experimental data demonstrate that the proposed algorithm can effectively eliminate the clutter which cannot be eliminated in the full aperture image, which eventually decreases the false alarm rate.