随机起伏海洋环境中的格林函数提取与应用

发布时间：2018-09-18 18:20

【摘要】：海洋环境噪声和散射现象所构成的背景场是引起声场随机起伏的重要原因。虽然传统上一般认为背景场是一种有害干扰,但是近几十年来应用背景场提取海洋环境信息的研究越来越多。本文研究的主要内容是背景场中格林函数的提取与应用。首先,本文研究了环境噪声干涉仪(格林函数恢复理论)。该理论表明利用两接收器处所接收的扩散噪声的互相关可以提取这两点之间的时域格林函数。针对实际海洋环境噪声的非扩散性,本文基于K-L变换和随机矩阵理论提出了一种新的数据预处理技术——扩散噪声滤波。实验数据表明该技术能够从非理想海洋环境中重构扩散噪声成分,增强格林函数恢复效果和被动测深仪输出响应。其次,本文将格林函数恢复理论和散射体成像技术结合起来开展了散射体被动成像的仿真和实验研究。针对格林函数矩阵恢复效果不理想而导致传统时反D.O.R.T和MUSIC等超分辨技术失效的情况,本文对克希霍夫偏移成像算法进行了改进,并利用压缩感知理论消除掉环境噪声互相关包络中的背景起伏,获得了良好的散射体被动成像实验结果。最后,根据以下原理——海底界面随机起伏和沉积层声速密度的非均匀分布可以引起测深仪回波的随机起伏,而不同海底底质类型(比如砂、粉砂、粘土等)的界面起伏和介质非均匀性具有不同的性质,因此导致的回波起伏特性也不同,应用一种时域高频后向散射模型,对黄渤海41个站位的20 kHz单波束测深仪回波数据进行了沉积层地声参数反演底质分类研究。采用下山单行模拟退火算法(SIMPSA)对模型输出和接收回波的平均包络曲线进行最大似然匹配,估计海底沉积层的平均颗粒度。实验数据处理表明：对接收回波包络进行能量归一化处理有效地减小了由实验不确定性所导致的回波幅度异常起伏,将反演参数的方差降低至未加处理的50%；平均颗粒度的反演最优值与沉积层样本实验室测量值的一致性更高；将反演结果用于底质分类,正确识别率达到75.6%。
[Abstract]:The background field formed by the noise and scattering phenomena in the marine environment is an important reason for the random fluctuation of the sound field. Although background field is traditionally regarded as a harmful interference, more and more researches have been done in recent decades to extract marine environmental information from background field. The main content of this paper is the extraction and application of Green's function in background field. Firstly, the environmental noise interferometer (Green's function recovery theory) is studied. The theory shows that the time domain Green's function between the two points can be extracted by using the cross-correlation of the diffusive noise received from the two receivers. Based on K-L transform and stochastic matrix theory, a new data preprocessing technique, diffusion noise filtering, is proposed for the non-diffusivity of the actual marine environmental noise. Experimental data show that the proposed technique can reconstruct diffusive noise components from non-ideal marine environments, enhance the recovery effect of Green's function and output response of passive sounder. Secondly, the simulation and experiment of passive imaging of scatterers are carried out by combining Green's function restoration theory and scatterer imaging technique. In view of the failure of traditional super-resolution techniques such as inverse D.O.R.T and MUSIC due to the unsatisfactory restoration effect of Green function matrix, the Kirchhoff migration imaging algorithm is improved in this paper. The background fluctuation in the cross-correlation envelope of ambient noise is eliminated by using the theory of compression perception, and good experimental results of passive imaging of scatterers are obtained. Finally, according to the principle that random fluctuation of seabed interface and non-uniform distribution of sonic velocity density of sediment layer can cause random fluctuation of echo of sounder, different types of seabed sediment (such as sand, silt, silt), The interface fluctuation and the medium inhomogeneity have different properties, so the echo fluctuation characteristics are different. A time domain high frequency backscattering model is used. The echo data of 20 kHz single beam sounder at 41 stations in the Yellow Sea and Bohai Sea are studied by means of sediment layer acoustic parameters inversion. Using downhill single row simulated annealing (SIMPSA) algorithm, the mean envelope curve of model output and received echo is matched with maximum likelihood, and the average particle size of submarine sediment is estimated. Experimental data processing shows that the energy normalization of the received echo envelope can effectively reduce the amplitude fluctuation of the echo caused by the experimental uncertainty and reduce the variance of the inversion parameters to 50 unprocessed. The inversion optimal value of average particle size is more consistent with the measured values of sediment samples in laboratory, and the inversion results are applied to the classification of sediment, and the correct recognition rate is 75.6.
【学位授予单位】：中国海洋大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P733.2

【相似文献】