复杂环境下的地基雷达大气改正方法

An Atmospheric Correction Method for Ground-Based Radar Under Complex Environment

地基雷达具有全天时、全天候、连续性和高时空分辨率的特点,能够对滑坡进行高精度近实时监测。但地基雷达形变测量易受大气干扰,特别是山区对流层时空变化影响强烈。通常大气变化被视为静态信号,利用距离函数模型进行改正。当发生强烈的大气湍流时,大气变化以动态大气信号为主,采用简单的函数模型将导致监测结果不可靠。借助通用型合成孔径雷达干涉测量大气改正在线系统中迭代对流层分解的成功经验,提出了一种地基迭代对流层分解(groundbased iterative tropospheric decomposition,GBITD)模型,在无全球导航卫星系统或其他任何外部天气观测值的条件下,将大气延迟迭代分解为分层分量和湍流分量,进而用于大气改正。这种方法具有很大的灵活性,将该方法应用于白格滑坡的滑后监测,以时间基线为2 min与10 min的两幅干涉图为例,其均方根分别从1.43 mm和1.69 mm减少至0.21 mm和0.24 mm。通过2018年12月4日至10日的778景影像得到在此期间白格滑坡最大累积位移为1400 mm,最大位移速率为230 mm/d。实验结果表明,GBITD模型比同质模型和分层模型的大气改正效果更加明显,可有效解决复杂山区利用地基雷达监测复杂的大气扰动问题。

Objectives:The ground-based radar has features such as working all-time in all-weather,providing continuous observations with high spatiotemporal resolutions and can monitor landslides with high precision in near real-time.However,atmospheric effects represent one major limitation of ground-based radar,due to the spatiotemporal variations of the troposphere,especially in mountainous areas.Atmospheric variations are often treated as static signals and corrected by the range function model which can be unreliable,particularly when strong dynamic atmospheric turbulence occurs.Methods:Based on its successful experience in the generic atmospheric correction online service for interferometric synthetic aperture radar(InSAR),the ground based iterative tropospheric decomposition(GBITD)model is utilized to decompose the tropospheric delay into stratification and turbulence components,and perform atmospheric correction.The GBITD model doesnt require global navigation satellite system or any external meteorological observations,which makes it flexible for InSAR atmospheric correction.Results:The application of the GBITD model to the 2018 Baige landslide suggests that the root mean square of two interferograms with temporal baselines of 2 min and 10 min decreased from 1.43 mm and 1.69 mm to 0.21 mm and 0.24 mm,respectively.The maximum cumulative displacement derived from the 778 images acquired during the period from 4 to 10 December 2018 was 1.4 m with the maximum displacement rate of 0.23 m/d.Conclusions:The experimental results show that the GBITD method performes best compared against the homogeneous model and the stratified model in complex mountainous areas.Furthermore,Ground-based radar has obvious unique advantages in landslide monitoring,which can provide strong technical support for landslide emergency monitoring and risk assessment.