海洋内波的多波束水体成像探测及其对测深影响的同步改正

Multi-beam echo sounding technology is currently the main technical means for high-resolution, high-precision submarine topographic surveying. The internal wave is a large-amplitude and high-velocity fluctuation phenomenon occurring in the seawater with stable density stratification, which is widely present in the ocean. When internal waves occur, the distribution of sound velocity in the ocean produces dramatical three-dimensional time-varying, which seriously affects the multi-beam sounding accuracy. The sounding data must be removed to meet the requirements of relevant standards. At present, there is no correction model that uses multi-beam to detect internal waves and synchronously corrects its influence on bathymetry. Therefore, this project is aimed at the influence mechanism of internal wave on the multi-beam sounding accuracy according to the internal wave dynamic propagation model. Based on multi-beam Water Column Image and submarine topographic features, the method of feature extraction and parameter inversion of the internal wave can be established. The obtained parameters then contribute to reconstruct the instantaneous shape of the internal wave, so as to solve the acoustic ray under the influence of internal waves, and the distortion correction of the seafloor topographic data is could be realized. The project will improve the quality of multi-beam sounding, facilitate the application of multi-beam sonar, and provide a precise and effective technical approach to marine internal wave research.

多波束测深技术是当前进行高分辨率、高精度海底地形探测的主要技术手段，海洋内波是发生在密度稳定层化的海水内部的一种大振幅、高流速的波动现象，其广泛存在于世界大洋。海洋内波发生时，海洋声速分布产生剧烈的三维时变，严重影响多波束测深精度，必须去除其影响后测深值才能满足相关标准要求。目前国际上还未出现利用多波束探测内波并同步改正测深影响的模型，为此，本项目以海洋内波动力学传播模型为基础，研究内波对多波束测深精度的影响机理，基于多波束水体影像和海底地形特征，建立海洋内波特征提取及参数反演方法，并根据得到的海洋内波参数重构内波瞬时形态，从而求解内波影响下的三维声线轨迹，实现海底地形数据畸变改正。项目将改善多波束测深质量，拓展多波束应用领域，为海洋内波研究提供一种精密且有效的技术手段。

项目主要针对海洋内波的多波束水体成像探测及其对测深影响的同步改正进行研究，拓展了多波束应用领域，提高了复杂环境下海底地形测量的精度。.基于海洋内波特征及动力学模型，分析了海洋内波对多波束测深精度的影响及海底地形畸变特征。分别从声速跃层垂直位移、声速跃层倾斜角度、航向方位角以及海底深度四个维度分析多波束测深中海洋内波的影响；针对不同因素影响机理，分析内波影响下的海底地形变化特点及测深误差分布，用以准确判断实测海底地形是否受内波影响。.针对受海洋内波影响的海底地形出现的类似“麻花”状褶皱异常，提出了基于多波束水体影像的海洋内波探测及同步水深改正方法。基于海洋内波的水体影像特征及受内波影响的海底地形畸变特征提取内波振幅、波长、传播方向等参数，构建内波瞬态方程；根据空间几何关系及Snell法则，确定声线折射路径，进行三维声线跟踪，计算波束脚印坐标；经过若干次迭代，逐渐提高改正精度。通过数值模拟表明，本文算法能有效改善多波束测深精度，实现海洋内波的探测及同步水深改正，提高多波束测量效率。.为削弱多波束水体影像中旁瓣噪声与背景噪声对目标的遮蔽，将深度学习方法引入水体影像的目标检测与分割，提出卷积注意力与金字塔池化结合的AP-UNet语义分割网络，实现多波束水体影像羽状流精细化分割。以UNet作为主体网络结构，利用卷积注意力模块和金字塔池化模块有效聚合多尺度上下文信息，实验结果表明，AP-UNet分割方法对复杂的水体噪声干扰有较好的抑制作用，适用于复杂水体环境下目标的自动提取与分割问题。.