机构地区: 广州地理研究所
出 处: 《热带地理》 2011年第4期388-392,共5页
摘 要: 时序差分干涉技术是近年来地面沉降监测领域的重要研究方向,基于时间序列SAR影像的稳定点目标识别和检测是进行时序差分干涉技术的关键和基础。本文针对短时间序列SAR影像的特点和普遍使用的点目标识别方法的局限性,借鉴散射体识别方法,综合考虑振幅和相位在时序SAR影像上的变化特征,采用多时相子孔径共轭内积和子孔径信息熵来获取时间序列过程中能够保持相位稳定的点目标,并用平均相干系数对选取的相干点目标进行修正。选取包含广州城区、东莞部分区域的2007年3月至2009年1月的17景Envisat ASAR数据进行实验,分别计算了方位向、距离向以及交叉向共轭内积和信息熵,并用基于高分辨率卫星影像数据解译的土地利用数据对提取的相干点目标进行验证。结果表明,利用多时相的子孔径影像共轭内积和信息熵可以得到比较可靠的相干点目标,交叉共轭内积在试验区域能获取精度高达92%的相干点目标。 Temporal differential interferometry technique has been one of the hot spots in land subsidence detection. It is proved that the coherent points in temporal SAR images are critical to the whole processing. Permanent scatter and Coherent point target are generally used to detect the coherent points, but they might pose problems to short temporal SAR images or low coherent area. In this paper, the temporal sub-look correlation method is used to derive the coherent stable points for SAR interferometry, which is validated with 17 scenes of Envisat ASAR images in Guangzhou city. Therein, sub-look correlation and entropy integrated with amplitude and phase information are calculated, and the coherent stable points are further corrected with average correlation among all the images used in this paper. The land use type is then used to validate the derive results. It is shown that a high accuracy 92% for coherent points can be achieved with the two-look internal Hermitian product ( IHP ) both in azimuth and range direction.