Back to results list
Show full item record
Please use this identifier to cite or link to this item:
|Title:||Integration of CGPS and ground meteorological observations for mitigating atmospheric effects on InSAR|
|Source:||Proceedings of ISPRS Commission II Symposium, Integrated Systems for Spatial Data Produc-tion, Custodian and Decision Support, Xi'an, China, 20-23 August 2002, p. 275-280 How to cite?|
|Abstract:||Repeat-pass interferometric synthetic aperture radar (InSAR) has been demonstrated useful for topographic mapping and surface deformation measurement. However, the interferometric phases of satellite radar signals are often seriously contaminated by atmospheric effects. Due to the highly variable nature of the atmosphere, especially the atmospheric water vapor, it is usually difficult to accurately model and correct the atmospheric effects. Consequently, significant errors are often resulted in InSAR measurements, especially in tropical regions like Hong Kong (HK).|
This paper studies a method for mitigating atmospheric effects on InSAR measurements based on an integrated use of Continuous GPS (CGPS) and ground meteorological observations. The tropospheric zenith delays (ZNDs) at a number of CGPS and ground meteorological stations in HK are first estimated and then used to construct a ZND map with a modified Kriging interpolator. Crossvalidation tests show that the mean value of the interpolation residuals is close to 0 and the RMS error is about 0.6 cm. When assuming a Gaussian distribution and under 95% confidence level, the interpolation errors are in range of –1.267 cm and 1.269 cm.
The study shows that the temporal and spatial variations of the tropospheric ZNDs can potentially cause a peak-to-peak error in a SAR interferogram of about 9.36 cm at the 95% confidence level for a one-day interval. The error increases to about 11.47cm for a ten-day interval. When the interpolation results are applied to correct the atmospheric effects, the peak-to-peak errors are reduced to 7.50 cm and 9.19 cm respectively for the one-day and ten-day intervals. This shows a 20% reduction in the peak-to-peak errors.
|Appears in Collections:||Conference Paper|
Show full item record
Checked on Oct 15, 2017
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.