摘要
高精度的高程测量需要通过干涉定标来对InSAR系统干涉参量误差进行校正,为了降低系统随机误差对定标精度的影响,一般选取GCP控制点时,基于敏感度矩阵条件数最小原则要求GCP控制点尽量沿整个测绘带均匀分布,但在复杂地形区域,由于野外布放GCP控制点难度较大,有时很难满足应用条件,因而针对基于区域网平差理论的稀疏控制点下InSAR定标方法中控制点的选取策略及误差传递关系进行分析。结果表明:选取GCP控制点时,如果不能满足整个测绘带内沿距离向均匀分布的条件,在满足尽量沿距离向分散的同时,GCP控制点位置沿距离向靠近远距端分布是一种较好的GCP布设方法。另外,基于区域网平差理论的多条带联合定标处理时,传递顺序应该由远距端向近距端依次进行。最后,对丘陵区域3条带仿真数据进行定标处理,通过对不同选取策略下各条带的定标结果以及不同传递顺序下3条带联合定标结果进行比较,验证了稀疏控制点下InSAR定标方法中控制点选取策略及误差传递关系分析的有效性及合理性。
High-precision height acquisition needs to correct the interferometric parameters errors by interferometric calibration for InSAR system.In order to reduce influence of the systematic random errors to the calibration precision,generally,while selecting GCPs,we follow the rule of minimum of conditional numbers of the sensitivity equations and choose GCPs along the whole strip of topography uniformly.but sometimes it's very difficult to satisfy this requirement as it's hard to dispose GCPs in the field with complicated terrains,so this paper makes researches on the strategy of GCPs select and error transfer manner among calibration method for airborne InSAR with sparse GCPs based on bundle adjustment,and analyzed results show that if we can't fulfill the condition that all the GCPs uniformly distribute along the whole mapping strip while selecting GCPs,trying our best to make the GCPs disperse along the slant range direction and putting them near the far end of slant range direction is a better way.Meanwhile,when we make a joint calibration of several strips,the processing order should be from the far end to the near end in turn.Finally,in comparison with results from different strategy of GCPs select and processing order,we make a calibration to the simulative data of three strips from the highland and the results confirm the validity and rationality of the proposed strategy of GCPs select and error transfer manner among calibration method for airborne InSAR with sparse GCPs based on bundle adjustment.
出处
《遥感技术与应用》
CSCD
北大核心
2011年第3期270-276,共7页
Remote Sensing Technology and Application
基金
国家自然科学基金项目(60890070)资助