复杂大场景下多极化SAR飞机目标检测关键技术研究

As one of the important tasks of battlefield sensing, the detection and recognition of radar ground targets has been a hot topic in the world. By providing much more terrain information than synthetic aperture radar (SAR), polarimetric SAR (PolSAR)attaches great importance to all countries in the world. Researchers invested much effort in the field of PolSAR application such as agriculture, forest, geology, hydrology and marrine monitoring etc. However there are very few research reports about the ground targets detection , especially the aircraft target detection based on PolSAR image under complex scene..This topic will focus on the key technical issues of the PolSAR aircraft target detection under complex scenes (including airplanes, vehicles, ships,airports,buildings, roads,railways,bridges, farmlands, oceans, rivers, forests and other targets in the scene).Joint sparse representation and optimization theory of multi-channel PolSAR image, nonlinear sparse classification theory, and polarization feature extraction method with rotational invariance will be applied in this topic to make a breakthrough in the aspects of speckle reduction of PolSAR, automatic extraction of airport area and airport runway, and detection of airplane targets etc. Our research will expand the PolSAR application field and provide new technical means for the real-time provision of military information and battlefield precision striking, which play an important role in the national defense construction.

雷达地面目标检测与识别是战场感知的重要任务之一，一直是世界各国的热点研究课题。极化合成孔径雷达（SAR）具有比单极化SAR 提供更为丰富的地物信息能力，受到世界各国学者高度重视，已在农业，森林，地质，水文，海洋监测等方面的应用研究取得了很多成果。但在利用极化SAR 图像进行地面目标检测特别是复杂场景中的飞机目标检测方面还很少有研究成果报道。.本课题针对复杂大场景下（含飞机、车辆，舰船，机场，建筑物，道路，铁路，桥梁，农田，海洋，河流，山林等目标）多极化SAR 飞机目标检测的关键技术展开研究。采用多通道复数域联合稀疏描述与快速优化理论、非线性组稀疏分类理论、旋转不变的极化特征提取方法，在极化SAR斑点抑制、机场跑道和机场区域自动提取，飞机目标检测等方面提出新方法并实现新的技术突破。对我国国防建设、扩大极化SAR 的应用具有重要意义，为实时提供军事情报、实现战场精确打击提供新的技术手段。