| As a relatively new technique, the Ground-Based Synthetic Aperture Radar(GBSAR) has gained interest as deformation measurement and monitoring tool in the last ten years. In the campaign, the instrument is installed in situ, acquiring data on a regular base every a few minutes. The GBSAR technique could offer high sensitivity to small displacements, in the region of sub-millimetres to millimetres, and long-range measurements, which can work up to some kilometres. The results acquired from this technology could support the research to the ground movement harzards. These features confer to the GBSAR technique a promising future.In practice, the GBSAR technique has to face two main problems. The first one is how to develop real-time capable technique which allows to analyze the GBSAR data and to assess the state of a mass movement with the least delay possible after a GBSAR measurement is acquired. The second one is how to overcome the limitation that the GBSAR technique could only extract the deformation information in the line of sight(LOS) direction. Firstly, In order to research the real time process algorithm, the paper focuses on three issues:(1) high quality Persistent Scatterer(PS) point seletion;(2) high precision interferometric phase error compensation;(3) PS real time deformation generation. Secondly, we introduce the multi-aperture interferometric theory(MAI) from the spaceborne DIn SAR process, and then investigate and apply it in the GBSAR field to extend the deformation measurement from one to two dimensions. The detailed contents and main contributions are as follows:(1) In general, the most used PS selection criteria are the coherence-based selection and the amplitude-dispersion-index-based selection. However, after applying the traditional method in the GBSAR real time process, the phase error of some selected PSs might be large or the PS set could hardly maintenance relative stabilization. Therefore, this paper presents the amplitude, coherence and phase joint PS selection algorithm to solve the problem. When considering the problem that the PS amplitude selection threshold varies in different campaigns, we apply the clutter distribution theory to estimate the clutter power in the SAR image, and then use the pixels’ signal to clutter ratio(SCR) information to determine the PS amplitude selection threshold adaptively. When facing the problem that the selected PS set changes with time during one campaign, we adjust the PS amplitude-dispersion threshold and amplitude threshold adaptively to keep the PS set relative stabilization. When solving the problem that the phase quality of PS varies with time, we make use of the coherence information and phase information to classify the selected PS set, and finally depress the negative influence induced by the low quality PS.(2) During the continuous mornitoing mode, as the influence of the nonideal factors, the repeat-pass error might occur. The phenomenon would lead to the fact that the zero-baseline condition could be hardly maintained. In order to overcome the difficulty that the repeat-pass error would degrade the deformation precision, the paper presents the algorithm which compensates the repeat-pass phase error and atmospheric phase error simultaneously. In our research, for the sake of establishing the fuction relationship between the repeat-pass error and its corresponding interferometric phase error, we firstly deduce transformation relationship between the repeat-pass error and the target’s phase error in SAR image, based on the repeat-pass error model. On this basis, we then analyze the influence from repeat-pass error to the interferometric phase error to establish the transformation model. After that, the least square(LS) theory is applied to compensate the interferometric phase error. Finally, the simulated data and real data were used to demonstrate the new method’s performance.(3) Considering the demand of real time monitoring during the continuous campaign, the paper present the GBSAR PS real time algorithm on the basis of the traditional spaceborne PS algorithm. When facing the problem of the process flowchart, we implement PS selection and PS deformation generation simultaneously, and divide the whole process period into two parts:(1) initialization period;(2) deformation generation period. When considering the problem that the PS number is not enough in the interferogram whose temporal baseline is too long, we divide SAR images into different groups and update the master image in each group. Meanwhile, the disappearance of some PSs might lead to the fact that the deformation information could not be extracted in some regions. Therefore, we apply the interpolation method to recover the deformation information in these regions. Finally, by analyzing the real data collected in the three experiments(Beijing Fangshan experiment, Heibei Tangshan experiment and Shanxi Lvliang experiment), we demonstrate the presented GBSAR PS real time algorithm is capable of acquiring deformation in the interest scene.(4) Considering the traditional GBSAR technique could only extract the deformation in the line of sight direction, the paper introduces the multi-aperture interferometry(MAI) technique from the spaceborne DIn SAR process, and then studies its application in the GBSAR field. In the spaceborne configuration, almost all targets are on the broadside geometry. However, in the GBSAR configuration, a variety of targets are on the squint geometry. Therefore, we first deduce the MAI algorithm on both two geometries in the GBSAR configuration. By comparison, we conclude that the MAI algorithm measures the deformation in the cross-LOS direction. Then, we discuss the MAI aperture selection problem after analyzing the deformation precision in the cross-LOS direction of MAI algorithm. Meanwhile, by applying the coherent sum method before the MAI process, the deformation precision is enhanced. Finally, the results of real data process show the MAI method has the ability to acquire deformation in the cross-LOS direction.
|
PDF Full Text Request