Improved Multichannel InSAR Height Reconstruction Method Based on Maximum Likelihood Estimation

Yuan Zhi-hui,Deng Yun-kai,Li Fei,Wang Yu,Liu Gang
DOI: https://doi.org/10.3724/sp.j.1146.2012.01530
2014-01-01
JOURNAL OF ELECTRONICS INFORMATION TECHNOLOGY
Abstract:In the application of getting the earth surface’s Digital Elevation Model (DEM) through InSAR technology, multichannel (multi-frequency or multi-baseline) InSAR technique can be employed to improve the mapping ability for complex areas with high slopes or strong height discontinuities, and solve the ambiguity problem which existed in the situation of single baseline. This paper compares the performance of Maxmum Likelihood (ML) estimation techniques with Maximum A Posteriori (MAP) estimation techniques, and adds two steps of bad pixels judgment and weighted filtering after the ML estimation. Bad pixels judgment is completed through cluster analysis and the relationship between adjacent pixels. A special weighted mean filter is used to remove the bad pixels. In this way, the advantage of the ML method’s good efficiency is kept, and the accuracy of DEM also is improved. Simulation results indicate that this method can not only keep good accuracy but also improve greatly the computation efficiency under the same condition, which is advantageous for processing large scale of data sets.
What problem does this paper attempt to address?