Line Array Pushbroom Images Digital Photogrammetric Processing

The geometric relationship of linear array push-broom imagery is stable and its quality is good. Using digital photogrammetric method, many kinds of important digital products can be obtained. In this thesis, the focus is put on the processing of linear array pushbroom imagery, which includes:1) the investigation on the exterior orientation, in which the estimation methods of initial value for exterior elements under different conditions are summarized, and a new orientationalgorithm called the robust combined ridge with shrunken estimator(RCRS) is presented;2) the disscussion of geometric rectification, in which the strict geometric rectification algorithm and many approximate rectification algorithms, such as direct linear transform(DLT), generic polynomial, improved polynomial and rational function model(RMF), are introduced in details;3) the research on the problem of bundle block adjustment, in which the strict linearization form of space intersection is deduced, and the basic principle and computation process of bundle block adjustment are discussed;4) the research on stereo matching problem, in which the epipolar rectification is studied, and then a structure information aided adaptive comprehensive matching algorithm(SIACM) is presented based on structure information aided multi-control and multi-criterion; and5) the elaboration of DEM generation, including DEM interpolation, the special processing for water area and vegetation, DEM filtering, error detection and accuracy check.Experiments results show that RCRS is a correct and accurate method for exterior orientation, and approximate rectification is a selective method for image rectification in some cases, and SIACM matching algorithm can effectively overcome the shortcomings of single information and criterion so that the reliable and accurate results can be achieved, and finally the reliable and accurate DEM can be generated using the methods presented above.