| Synthetic aperture radar interferometry (InSAR for short) is a newly developed spaceremote sensing technology; it is the combination product of the conventional SAR remotesensing and radio astronomy interferometry technology. Interferometric phase image can beextracted through joint treatment for two radar images of the same area, and then digitalelevation model (DEM) can be established from it. SAR imaging has many advantages,such as all-weather continuous observing capabilities, hight-resolution, high-precisionimaging, wide coverage and low cost. This technology is widely used in many fields ofnational economy, like topographic mapping, geological research, disaster prevention andmitigation, agricultural and forestry, and marine research etc.Doris software is developed by the Dutch Delft University of Technology in1999; it isa set of open source, cross-platform InSAR processing software. The software is able tocompletely implement the entire procedure of InSAR data processing, and optionalalgorithms are provided for multiple processing steps. At present, it can handle variousspaceborne radar data, such as the ERS-1/2and ENVISAT data of ESA, the Japanese JERSand ALOS data, and Canadian RADARSAT-1data, etc. Although Doris software has strongprocessing ability, and wide range of application, its conveniences of user visualizationoperation is incompatible with the current software system under window environment. Fornon-professionals in the field of remote sensing, the complex way of param setting as wellas file modification, and the command line type of data processing is “difficult to install,slow to master, verbose to operate, and strongly professional”, which makes it easy togenerate operational issue caused by personal factor. Thus, the processing result isunsatisfactory, or even has mistakes.Phase unwrapping is a key step in InSAR data processing. Whether the phaseunwrapping algorithm is practical and can meet the precision requirements directlyinfluence the establishment of digital elevation model. A network flow algorithmSnaphu is proposed by Chen in2001, which takes better computational efficiency andaccuracy of the two best results into account. However, through the course ofunwrapping the filtered interferogram for rugged area using Snaphu algorighm, there are still two existing problems:(1) the unwrapping process do not always can get results;(2)the unwrapping result is highly influenced by data coherence factor. In particular, fordense vegetation and humid air regions, decorrelation factors are more serious. This isundoubtedly hindered the application of InSAR technology in rugged terrain.A convenient and friendly SAR data processing system under visualizationoperating environment—VDoris (Visual Doris) system, whose kernel is Doris system, isset up for non-professionals in the field of remote sensing, which is based on certaindeficiencies of the existing Doris software. The graphical user interface is the operatingappearance while the functional module is controlled by business logic management. In thisstudy, firstly the architecture and the data processing flow of Doris software is analysised.32subfunctions and33parameter configuration files are splited by means of unlinking thefunctional modules and parameter configuration files of Doris software.3parameter typesare summarized through parameter analysis and classify. In the meantime, combined withthe characteristics of object-oriented programming, a module class is put forwarded to bethe base class of all the subfunctions via analysising the mutual operations and attributes ofall the subfunctions, which makes it has better stability, reusability, and maintainability.Finally, a set of Qt-based visual VDoris system under Linux is developed. This system hassuch characteristics as full-featured, friendly interface, multi-modes, smart and simple, andso on, while containing the full functionality of SAR data processing as well as offeringoneclick processing tools.In the aspect of phase unwrapping algorithm research, a “mask approach” is proposedto improve the Snaphu algorithm. Through this approach, low-coherence pixels are shielded,and do not participate in the unwrapping operation, so as to control the error propagation,and promote the unwrapping velocity. A secondary treating like surface fitting isimplemented to the unwrapped result in order to fill the vacancy points, which makes theunwrapped phase values smoother. At last, the improvement of unwrapping algorithm isanalysised and proved by means of comparing the generated digital elevation model withthe reference DEM. |
PDF Full Text Request