一种精细化声速场分步构建方法  

A stepwise construction method for refined sound velocity field

作　　者：刘子阳 王振杰[1] 王兆喆 LIU Ziyang;WANG Zhenjie;WANG Zhaozhe(College of Oceanography and Space Informatics,China University of Petroleum(East China),Qingdao 266580,China)

机构地区：[1]中国石油大学(华东)海洋与空间信息学院,山东青岛266580

出　　处：《测绘工程》2025年第1期30-38,共9页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(42174020)。

摘　　要：海洋环境监测、水下导航定位都需要准确的声速场信息,而声速场的构建面临分辨率与工作范围、时空尺度与数据之间的矛盾,导致无法反映准确的声速变化。为此,本文将声速场分为大尺度时空分辨率较低的背景场和小尺度时空分辨率较高的扰动场,提出一种精细化声速场分步构建方法。首先,通过Argo声速剖面数据,基于经验正交函数构建声速长时间物理空间变化的背景场;然后,通过实测声速剖面数据,基于人工蜂群优化BP神经网络构建声速瞬时动态变化的扰动场;最后,将背景场与扰动场结合构建精确的声速场模型,反映完整的声速变化信息。本文基于该方法构建了南海局部区域600 m水深的声速场,结果表明,该方法能够建立高精度的声速场,其预测精度可达到1.038 m·s^(-1)。The accurate sound velocity field information is needed for marine environmental monitoring and underwater navigation and positioning.However,the construction of sound velocity field is faced with the contradiction between resolution and working range,spatio-temporal scale and data,which leads to the inability to reflect the accurate change of sound velocity.Therefore,this paper divides the sound velocity field into the background field with low spatio-temporal resolution at large scale and the perturbation field with high spatio-temporal resolution at small scale,and proposes a stepwise construction method for refined sound velocity field.Firstly,the background field of long-term physical space change of sound velocity is constructed based on the empirical orthogonal function through Argo sound velocity profile data.Then,the perturbation field of in-stantaneous dynamic change of sound velocity is constructed based on the artificial bee colony optimization BP neural network algorithm using the measured sound velocity profile data.Finally,the background field and perturbation field are combined to construct an accurate sound velocity field model,reflecting the complete change of sound velocity.This paper constructs the sound velocity field at the depth of 600 meters in the local area of the South China Sea based on this method,the results show that this method can build the high-precision sound velocity field,and its prediction accuracy can reach 1.038 m·s^(-1).

关 键 词：声速场 声速背景场 声速扰动场 经验正交函数 人工蜂群优化BP神经网络 

分 类 号：P229[天文地球—大地测量学与测量工程]