Research On Underwater 3D Reconstruction Scheme Based On 2D Imaging Sonar

Recovering 3D information from multiple frames of 2D sonar images is the core problem of 3D reconstruction of underwater acoustics.There are many sonar devices used for 3D reconstruction of underwater acoustics.Different sonar devices have different physical characteristics.The differentiation of sonar devices in specific needs for reconstruction algorithms is also proposed..This thesis mainly studies the 3D reconstruction algorithm for two kinds of sonar equipment,wide-aperture forward-looking sonar and side-scan sonar.The results of the thesis are as follows:(1)Firstly,a 3D reconstruction method based on multi-frame image of forward-looking sonar is proposed.Thesis expounded the sonar image equation and the meaning of blind deconvolution,discussed the specific range of constraints that need to be considered in the actual solution process,analyzed the influence of constraint parameters on the solution vector,and the direction of parameter setting.In the experiment,several underwater objects were reconstructed,the three-dimensional recovery ability of the algorithm for small-scale underwater targets was verified,and the visual representation of small-scale underwater targets was obtained.(2)Another algorithm for reconstructing and restoring seabed elevation from side-scan sonar images and several known elevations is pointed out.On this basis,iterative process be improved.Introducing the reconstructed elevation error into the algorithm in each iteration by trig transformation,and make certain constraints on the mechanism of iterative non-convergence.The closed-loop effect introduced by the error is verified by simulation,and the convergence and rapidity of the reconstruction iteration are also verified by the simulation.Finally,the threedimensional elevation value and seabed reflectance value of the seabed under the simulation situation are restored visually.(3)And then focuses on the height sounding requirements of the side-scan sonar reconstruction algorithm.A time-stamped synchronous three-array short-baseline shallow sea acoustic localization algorithm based on the TOA(Time of Arrival)principle is proposed.Based on this algorithm,and develop a set of underwater acoustic localization system for calibrating the coordinates of side-scan towfish.This thesis gives a more complete introduction to the software and hardware of our positioning system.Three simple algorithms are proposed to solve the problem of elevation point acquisition in side scan sonar reconstruction algorithm.(4)Finally verified the visualization ability of the busy deconvolution target recovery algorithm based on wide-aperture forward-looking sonar through the actual water tank test.Afterwards,Then collected real side-scan sonogram data through sea trials and lake trials,and verified the 3D recovery capability of the seabed reconstruction algorithm based on side-scan sonar image and elevation prior.Finally,through the offshore test,the positioning accuracy of the short-baseline positioning system under static and dynamic conditions and the hardware stability of the positioning system are evaluated and calculated.Experiments show that our underwater acoustic positioning system has achieved a certain accuracy and has the ability to quickly deploy in shallow seas.