基于二维侧扫声纳图像的三维高分辨率海床地形恢复方法研究

High-resolution seabed topography takes a basic guarantee in marine project, exploitation of ocean resource, etc. High-resolution sonar image can be acquired by side-scan sonar system (SSS), which can clearly show seabed relief, but is short of depth information in it. It has very-important theoretical and practical significance to develop the theories and methods of the recovery of 3-D high-resolution seabed topography with 2-D SSS image, which can compensate the disadvantage of traditional SSS image, and produce a new technique for acquiring high-quality seabed topography as well. .Based on the pre-processing of SSS observation, the theories and methods of sound-wave reflection on seabed are studied firstly. Both of the synthetic sound-wave reflection model and the restriction model of seabed-topography recovery for SSS are given out, which are the base of seabed-shape recovery. Then, based on the acoustic theory, visual theory and the theory of SSS image forming, the methods of seabed-shape recovery are developed. In the work, the optimum seabed-recovery algorithms are studied and come up with under different restriction conditions, and the synthetic model of seabed-shape recovery is also presented. By means of these algorithms and models, high-resolution seabed shape is achieved. Finally, the methods of constrained control are studied. The synthetic algorithms of constrained control and the evaluation method are put forward, which can guarantee the quality of the recovery, and at the same time introduce the absolute vertical datum into the recovery. Based on above, the high-precision and high-resolution 3-D seabed topography is achieved. .The theories and methods proposed in the research make a promotion in digging the potential application of traditional SSS image, completing the theories of acquiring high-quality seabed topography, developing new investigation instruments, etc.

高质量海床地形对于实施海洋工程、开发利用海洋资源等具有非常重要的基础保障作用。传统侧扫声纳可获取高分辨率海床地貌图像，但缺少水深信息。基于二维侧扫声纳图像，开展三维海床地形恢复理论和方法研究，对于弥补侧扫声纳测量成果的不足，形成一种全新的高质量海床地形信息获取技术，具有十分重要的理论和现实意义。项目首先开展声波在海底的反射理论和方法研究，提出并构建针对侧扫声纳的海底声波综合反射模型和海床形状恢复约束模型；其次，开展海床形状恢复方法研究，给出不同条件下的恢复算法，构建最优综合恢复模型，获取高分辨率海床形状；在上述基础上，开展约束控制方法研究，给出恢复海床形状的约束控制算法及精度评估方法，形成高精度、高分辨率海床地形。项目研究提出的理论和方法对于目前广泛应用的传统侧扫声纳系统测量成果应用潜力的挖掘、高质量海床地形信息获取理论和方法的完善、新型海洋扫测系统的研制和应用等具有促进作用。

针对侧扫声呐（SSS）图像在表达海底地形起伏变化方面的空白，开展了基于二维（2D）SSS图像的三维（3D）高分辨率海底地形恢复理论和方法研究，提出了一种顾及底质分布的辐射畸变改正方法、SSS图像海底线综合跟踪方法、顾及共视目标的多条带图像拼接方法、顾及镜反射和漫反射的海底综合反射模型、基于SFS的2D声呐图像反演3D海底地形方法及最优反演模型、基准和尺度约束模型构建方法及基于外部测深数据约束的3D绝对海床地形恢复方法、一种改进的线性反演模型及3D绝对海床地形恢复方法、一种通过挖掘SSS海底线、无需外部测深数据约束的3D海床地形恢复方法，并研制了基于SSS 图像的3D海床地形反演软件，实现了在有和无外部测深数据约束下的3D高分辨率海床绝对地形恢复。研究取得了与测深相同的地形反演精度和50~70倍测深分辨率的海底地形，已发表论文34篇，获批软件著作权1项，参加5次国际学术会议及9次主要国内学术会议，已培养博士和硕士研究生10名。 .项目研究达到了预期目标，扩展了现有SSS的应用潜力，为高分辨率海床地形的获取提供了一种全新的途径，为我国新型测深设备的研制提供了理论和技术支撑。