Research On Underwater SAS Data Modelling Method And Key Technology Of Interferometry

Since starting of the21st century, there is larger and larger step of human exploration plans for using the marine resource, which necessitate information acquisition technology in the submarine world to assists human ocean exploration plan, and being of extremely important practical values. Synthetic aperture sonar (SAS) has profited from success application of the synthesis aperture radar (SAR) technology, overcoming contradictory relations working range and the resolution of image on traditional side scaning sonar and, archiving range independent high resolution of sonar ranging formation technology in theory. More profound research on SAS technology is of extremely vital significance for advancing underwater survey and mapping technology, submarine remote sensing and underwater topography survey.Existing research on SAS, mostly concentrates in on several aspects, including SAS data simulating, motion compensation, SAS imaging algorithm, SAS interferometry and SAS image processing. Among them, the data simulating method mainly by software way, according to submarine sound wave propagating principle and scattering characteristic, simulate real hardware system data acquisiting procedure. The SAS data simulation method, avoid high cost and the low efficiency of practical data aqcquisition, providing the ideal experimental data for the relative research. The SAS interferometry (InSAS), acquisiting SAS image from vertical installment receiver array and coherent processing, obtaining relative time delay observing target deriving from corresponding interferogram, then terrain information of observing target is available. The InSAS technology take advantage of high resolution SAS data, thus bring out high resolution underwater topography data.This thesis mainly aims at the underwater topography, concerning SAS data modelling method and key technologies of SAS interferometry, the main research content contains following:(1) SAS data simulation method based on DEM model. Existing simulators of side scaning sonar, are mostly based on the narrow beam and narrow band width pattern, and SAS system is of wide beam and wide band signal, moreover the SAS data depend on coherence of data, needing phase information of data, it is not meet for the traditional side scaning sonar simulator. In view of submarine virtual scene, descripting mode of topographical model is an important factor forhight processing efficiency for SAS data simulation, using rough facet model is possible effective to avoid a big computation burden, and moreover archive a higher simulation compared to the smooth surface model. Underwater the sound wave backscattering of floor has the directivity, moreover havesomething to do with under water material attribute. In addition, signal masking shadow region caused by fluctuating of underwater topography, is also another important factor for reality of simulating data. In the thesis, considering efficiency and reality of simulateing result, reviewing and analizing necessary aspact of SAS echo data and the SAS image data simulating methods.(2) Proposing detecting and matching method for SIFT feature of SAS the image and the image registration method based on optical flows. Because the underwater environment SAS data acquisiting is complex, and sonar image is usually polluted by noise severely, causing diffucity in extracting matching characteristic from image. The SIFT mathing method, locating features by gray spatial distribution characteristic of image data own. Its feature extracting has pointed, and extracted features carry on the full descriping information. Feature matching based on SIFT point feature by distance of descriptor make higher matching precision variable. Optical flow methods adaptive to non-rigid distorts for SAS image registration, make use of characteristic of good coherence and small offset carries quantity on the SAS image pairs. Its non-rigid deformation descripting ability is more flexible to displacement well which fluctuatation of underwater floor. By multi-criterion strategy, optical flow method existence and prior terrain based method could avoid partial minimum problem.(3) Proposing a new minimal norm phase unwrapping method. Based on analysis of the existing minimal norm phase unwrapping methods, here concentrate in formulation of the cost function. Because the classical L2norm cannot obtain the quite ideal result for unwrapping, and the most phase unwrapping method obey to the cost function of L0and L1norm whose classical definition is underivable, not meet variation method requests. Through new cost function is provided, obtaining new Lp norm Euler-Lagrange equation, by two simple parameter, tranditional methos are included.