浅海地声参数的变维贝叶斯反演  

Trans-dimensional Bayesian geoacoustic inversion in shallow water

作　　者：吴伟文 任群言 鹿力成[1,2] 郭圣明 马力[1,2] WU Weiwen;REN Qunyan;LU Licheng;GUO Shengming;MA Li(Key Laboratory of Underwater Acoustic Environment,Institute of Acoustics,Chinese Academy of Sciences Beijing,100190;University of Chinese Academy of Sciences,Beijing,100049)

机构地区：[1]中国科学院水声环境特性重点实验室(声学研究所),北京100190 [2]中国科学院大学,北京100049

出　　处：《声学学报》2023年第3期496-505,共10页Acta Acustica

基　　金：国家自然科学基金项目(12274431);中国科学院声学研究所自主部署项目(MBDX202102)资助。

摘　　要：为解决海底沉积层分层结构未知时的反演问题,提出了一种变维粒子滤波方法,利用声场的互谱密度,估计沉积层分层结构以及地声参数。仿真结果表明:变维粒子滤波在沉积层分层结构未知时,能有效反演沉积层层数以及地声参数,粒子的并行计算能使其相较于可逆跳蒙特卡洛马尔可夫链(rjMCMC)更加高效。利用变维粒子滤波,对南海垂直线阵列接收到的线性调频信号进行处理,反演结果与rjMCMC反演得到的沉积层层数和地声参数结果相近,说明了此方法能有效估计沉积层层数的同时反演浅海地声参数,得到可靠的参数后验概率密度。To deal with the inversion problem when the spatial structure of seafloor sediments is unknown,a trans-dimensional particle filter method is proposed in this paper,where the cross-spectral density of the pressure field is used to estimate the sediment layering structure and geoacoustic parameters.The simulation results show that the number of sediment layers and geoacoustic parameters can be effectively estimated using the proposed method,and the parallel particle calculation make this method more efficient than reversible jump Monte Carlo Markov chain(rjMCMC)sampling.A linear frequency-modulated signal received by a vertical line array in the South China Sea is processed using the proposed trans-dimensional particle filter.The inversion results of the sediment layers and geoacoustic parameters are similar to those acquired by rjMCMC inversion.The number of sediment layers and the posterior probability density of parameters can be effectively estimated using this method.

关 键 词：地声反演 地声模型 贝叶斯理论 不确定性分析 

分 类 号：P733.2[天文地球—物理海洋学]