基于多源水深数据融合的海底高精度地形重建

Bathymetry fusion techniques for high-resolution digital bathymetric modeling

本文在研究多源水深数据构建技术的基础上,分析了张力样条插值算法和"移去-恢复"法的多源水深数据融合处理技术,基于该方法选取实验区,利用多波束、单波束、历史海图等多源水深数据进行高精度海底地形融合试验,并针对多源水深融合技术缺少误差评估的现状,利用split-sample方法对融合结果进行水深不确定性评估,形成融合结果的可靠性空间分布。结果表明该方法无论是在数据稀疏区还是高密度区都达到了较好的融合效果,既保留了高分辨率水深数据的细节信息,又较真实的反映了研究区海底地形特征,且构建的海底地形精度可靠,误差百分比集中在0.5%。本文整套数据融合和结果评估方法可为多源水深数据融合的海底高精度地形构建提供借鉴和参考。

This paper reviews current fusion techniques used for bathymetry data,introduces the＂Splines-in-Tension＂interpolation and＂remove-restore＂techniques for bathymetry fusion.Based on these techniques,high-resolution digital bathymetric model is built using multi-beam,single-beam and historical charts data sets.In addition,facing the lack of error estimates for the interpolated grid points,the paper uses split-sample method to assess the uncertainty of the result for reliability.It shows a satisfied result with error percent about 0.5%both in sparse and dense areas,preserving the detail information of high-resolution data,possessing the vivid characteristics of the sea bottom.The whole techniques can be applied in multi-sources data fusion and combination.