Research On DEM/DOM Generation Method For Airborne Millimeter-Wave InSAR

In the cloudy and rainy areas such as southwestern China,when DEM and DOM are acquired by means of optical remote sensing and photogrammetry,it is difficult to obtain optical images due to the influence of cloud,fog,rain and other weather conditions,resulting in data missing in some areas.InSAR is not affected by weather conditions,and it can acquire large scale data full time and full weather without resolution loss in distance,which is one of the significant means for obtaining DEM.In particular,the airborne dual-antenna millimeter-wave InSAR system at Ka-band can acquire DEM over difficult mountainous areas with steep or complicated topography efficiently and accurately for its short wavelengths,flexible maneuverability,variable track,high coherence,high resolution,real-time acquisition,etc.,which reduce the difficulty of acquiring DEM in complicated areas.In China,there are few research institutes of airborne dual-antenna millimeter-wave InSAR and there are few experiments of extracting DEM / DOM from topographic mapping.In this paper,the key technologies of generating DEM / DOM from airborne millimeterwave InSAR are studied.The main research works are as follows:(1)Systematic interference parameters affecting DEM accuracy are obtained from the 3D reconstruction model,and the sensitivity of these parameters is analyzed.According to the sensitivity analysis results,the GCPs are used to calibrate the interference parameters of the three system parameters to obtain the corrected system parameters,thus eliminating the influence of residual errors.(2)The interference parameter calibration method calibrates the system parameters based on GCPs.For large-scale,multi-strip or areas(mountains,hills,swamps,etc.)where GCPs are not easy to be arranged,this paper uses a small amount of GCPs and a large number of selected TPs to establish the constraint relation between images,and accurately corrects the system interference parameters of each interference image pair according to the block adjustment algorithm,while solving the coordinates of the TPs in the measurement area.(3)The DSM is prefiltered by using a filtering algorithm combining median filtering and morphological open operation to filter out elevation anomalies.Since DSM includes buildings,bridges,trees and other ground objects higher than the ground,in order to obtain DEM,these ground objects need to be removed,this paper proposes a filtering method to detect ground objects from four directions,filter out ground objects and fill them to obtain corresponding DEM.