Research And Implementation On Key Techniques For Deriving DEM From Airborne InSAR

Synthetic aperture radar interferometry(InSAR) is the earth observation technology deriving ground elevation from phase information, extensively applied in mapping, ocean, Polar region and so on, makes great contribution to national security and economic development. It is an important research direction of InSAR to derive DEM. This paper mainly study and implement key technology of airborne InSAR data processing method including complex image co-registration of airborne InSAR, interferogram flattening, phase unwrapping, phase to height conversion, Geocoding, develops Airborne InSAR data processing system,promotes the development and application of InSAR technology.The main contents of this thesis are as following:(1) We study complex image co-registration process of airborne InSAR and introduce four common complex image co-registration methods:real correlation coefficient method, complex correlation coefficient method, average fluctuation function method and maximum spectrum estimate method. For adaptability of co-registration measures in different terrain is different, we use four kinds of co-registration measures to test the data of airborne InSAR mountain and flat land separately and the accuracy and effectiveness of four methods are compared and analyzed. It is concluded that maximum spectrum estimate method can get better registration result in the flat land, complex correlation coefficient method can obtain a good registration result in flat land and mountain land, which is the best of four registration method of InSAR complex image. In addtion, the strategies of partition processing and OpenMP acceleration for large image are studied to improve the ability and efficiency of large image processing.(2) We study and implement interferogram flattening method based on orbit, interferogram flattening method based on coarse precision DEM, vector filter'method and main frequencies extraction filtermethod firstly. Besides, we summarize the existing methods of phase unwrapping, and analyze the principles of all kinds of methods, finally introduce principles and implementation steps of classical branch cut method and region-growing method and validity of two algorithms are compared and analyzed by experiments.(3) We study phase to height conversion and geocoding, and the feasibility of algorithms is verified by experiments.(4) We take VS2010 as development platform, combines with C++language to achieve algorithms from the third chapter to the sixth chapter and develops airborne InSAR data processing system that includes scheduling management module, interference processing module and data display module. It can generate DEM by interferencing processing using InSAR data. Finally the precision of product is verified by airborne InSAR measured data, the error in the X direction is 1.080m, the error in the Y direction is 1.055m,and the error in the H direction is 0.822m.