Seafloor Classification of Multibeam Sonar Data Using Neural Network Approach

In this study, the self-organizing map (SOM), which is an unsupervised clustering algorithm, and a supervised proportional learning vector quantization (PLVQ), are employed to develop a combined method of seafloor classification using multibeam sonar backscatter data. The PLVQ is a generalized learning vector quantization based on the proportional learning law (PLL). The proposed method was evaluated in an area where there are four types of sediments. The results show that the performance of the proposed method is better than the SOM and a statistical classification method.     

最大似然分类法在多波束底质分类中的应用

研究利用多波束测深系统获取的反向散射强度数据和海底声像图,通过数据预处理,应用Beyes统计分类方法实现了对海底泥、砂、砾石和基岩等底质类型的自动分类识别。     

基于改进BP神经网络的海底底质分类

通过采用遗传算法优化神经网络初始权值的方法,将GA算法与BP神经网络有机结合,应用于海底底质分类。基于多波束测深系统获取的反向散射强度数据,应用改进的BP神经网络分类方法,实现对海底基岩、砾石、砂、细砂和泥等底质类型的快速、准确识别。通过实验比较,GA-BP神经网络分类精度明显高于BP神经网络,证明了该方法的有效性和可靠性。     

多波束海底地形图的制作方法探讨

提出了采用高斯距离函数加权平均算法,对离散多波束数据进行网格化,该算法的优点是权函数收敛较快,能较好地保留细小的地形特征,而且运算速度较快。对于图形制作方面,提出了2D等深线图叠加3D阴影图模式的海底地形图的绘制思路及实现方法,最终成果图立体感强,可以更好地表现出地形细节。最后就如何改进海底地形图的配色方案进行了探讨。     

Operational Parameters,Data Density and Benthic Ecology: Considerations for Image-Based Classification of Multibeam Backscatter

Efforts to develop a procedurally robust method for automated classification of multibeam backscatter have taken a variety of approaches (e.g., image-based, textural, angular range analysis). For image-based classification, little research has focused on the roles of operational parameters of vessel and sonar system in affecting the final classification. Repeat multibeam surveys (2005 and 2006) conducted at the same area with different sounding densities were classified using QTC-Multiview. Comparison of class areas revealed 78% agreement between classifications derived from the two surveys. Cross-tabulation of ground truth video and class demonstrate 71% agreement in the low-density survey and 77% for the high-density. Differences between classifications are primarily attributed to variation in along track data density, errors in the compensation process, and/ or insufficient quality control of the input data. Natural change detection at the scales observed was determined not to be practically discernable from the errors associated with the classification process.     

多波束回声数据的统计与底质分类应用

利用声反向散射数据作海底沉积物分类,是海洋地质学家感兴趣的话题,也是目前多波束声纳应用的一个研究热点。结合胶州湾实际调查数据,探讨了贝叶斯分类方法在该领域的应用。研究结果表明,该方法可以对不同的底质类型进行分类,可以识别未知的底质类型以及对混合在一起的两种不同类型的目标进行分类。     

多波束测深海底数字地形模型的建立

针对多波束测深数据集,采用改进的距离反比权重算法和多细节层次模型技术来建立海底数字地形模型（DTM）。结果表明：本方法具有较高的精度和效率,可以满足不同分辨率要求的DTM应用。     

自组织数据地图在大辽河河口水质分析中的应用

自组织数据地图(SOM,Self-organizing Map)具有强大的数据压缩和提炼功能,应用它可以对水质监测数据进行处理,找到水质分析的新方法。本文以大辽河河口水质监测数据为例,应用MATLAB中SOM工具箱,采用批处理SOM算法,利用MATLAB语言编写出水质分析程序。该程序通过SOM的训练对数据特征进行分析和显示,并进一步分析SOM训练结果,将监测数据分成4类,最后实现数据跟踪及新的监测数据的自动归类。该方法可以将水质监测数据的特征可视化,便于直观、深入地分析水质状况,为识别水质特征、跟踪水质变化和应对措施提供决策支持。     

海底沉积物声学特征定量分析及其智能分类研究

海底底质特性描述及分类是当今浅海声学的研究热点,海底沉积物的物理结构特性与其声学响应特征密切相关。在分析海底沉积物声传播特性的基础上,应用现代计算机信号分析技术手段,对海底沉积物声学响应波形提取了4个特征参数:声速、波幅指数、波形关联维分形指数和声波频谱的频率矩。以这4个特征参数作为输入向量,海底沉积物的结构类型作为输出向量,建立径向基概率神经网络模型。研究表明建立的神经网络模型具有较强的海底沉积物分类预报能力。     

基于自组织映射的日本南部黑潮与黑潮延伸体的典型时空模态及其因果关系研究

过去的研究认为,黑潮延伸体的年代际振荡受来自其下游的太平洋年代际振荡（PDO）相关联的信号主导,但最近的观测表明这种调控机制在2017年9月之后不再成立。与此同时,黑潮延伸体的上游即日本南部黑潮正在发生一次大弯曲事件。利用26年（1993–2018年）的卫星高度计提供的海表高度距平数据和自组织映射（SOM）方法,本文研究了日本南部黑潮与黑潮延伸体的时空模态及其因果关系。结果表明,SOM能有效地提取两个海区的典型空间模态,且它们的演变轨迹表明当日本南部黑潮处于大弯曲（离岸型非大弯曲）路径时,黑潮延伸体趋于稳定（不稳定）态。基于SOM识别得到的海表面高度距平（SLA）特征区及特征时间模态,我们进一步利用一种最近发展的定量因果分析方法研究了两个流系之间的因果关系。研究发现,当黑潮大弯曲发生时,日本南部黑潮和黑潮延伸体之间存在双向因果,但因果关键区不同。前者对后者的影响集中在纪伊半岛东南侧及黑潮延伸体“两脊一槽”区域,而后者对前者的影响则集中在黑潮延伸体“两脊一槽”区域及黑潮再循环流区域。这说明黑潮大弯曲的发展对黑潮延伸体的稳定性有重要作用,同时黑潮延伸体通过调制南部再循环流影响日本南部黑潮的路径。不同的是,当离岸型非大弯曲路径发生时,只有从日本南部黑潮向黑潮延伸体的单向因果关系,且因果性主要集中在伊豆海脊及再循环流区域。这与该时期海表高度负异常沿日本南岸不断向位于下游的黑潮延伸体再循环流的传播有关,它使得黑潮延伸体变得不稳定。     

面向分类的高光谱遥感影像数据特性的研究

分析了高光谱遥感的特点、影像数据的应用现状,针对高光谱遥感影像处理、分类的现状及遇到的困难,对高光谱遥感影像数据的分类特性进行了深入的分析总结,对高光谱遥感影像分类的支持向量机方法进行了系统研究,并分析指出了支持向量机分类法的优越性,最后提出了需要进一步研究的问题.     

Development of a Vector-Based Method for Coastal Bluffline Mapping Using LiDAR Data and a Comparison Study in the Area of Lake Erie

This paper introduces a new method for bluffline extraction based on analysis of surface normal vectors derived from a 2.5-dimensional Delaunay triangulation network generated using a set of LiDAR points. The developed method has been implemented and tested for the extraction of blufflines in a section of the Lake Erie coastline near Painesville, Ohio. Comparison of the research results with those generated by a method for bluffline extraction based on slope information along elevation profiles on transects shows an improvement in both horizontal and vertical accuracies in situations where the bluff has a relatively sharp edge formation.     

An Advanced Probabilistic Neural Network for the Design of Breakwater Armor Blocks

In this study,an advanced probabilistic neural network(APNN)method is proposed to reflect the global probability density function(PDF)by summing up the heterogeneous local PDF which is automatically determined in the individual standard deviation of variables.The APNN is applied to predict the stability number of armor blocks of breakwaters using the experimental data of van der Meer,and the estimated results of the APNN are compared with those of an empirical formula and a previous artificial neural network(ANN)model.The APNN shows better results in predicting the stability number of armor blocks of breakwater and it provided the promising probabilistic viewpoints by using the individual standard deviation in a variable.