基于多波束数据的东海陆坡区地形分类

【作者】 陈义兰；

【导师】 刘忠臣；

【作者基本信息】 国家海洋局第一海洋研究所 ， 环境科学， 2007， 硕士

【摘要】 海底地形分类是一种分析海底地形发育特征的复杂程度、将海底形态划分为不同的地形单元的地形描述方法。它是对海底地形信息的一种表达分析方法,是海底地貌学研究的重要内容。目前对海底地形的分类主要限于描述性的人工分类,大大降低了地形信息的可靠性。地形的自动化分类是数字化地形分析的精确、直观的表现形式,是对地形的一种定量描述,而这种定量、精确、客观的描述将为海底地貌综合制图提供依据,为海底地质单元的划分和地质构造的识别提供重要的参考价值,为海洋灾害研究和油气资源勘探提供重要的基础地形信息,但是目前在这方面的研究还相对较弱。东海陆坡地形复杂,是许多领域研究的重点区域。对坡折线、坡脚线等地形特征的客观、准确确定,对东海海底地学研究具有很深远的科学意义和实际应用价值。东海的大部分海域已经积累了高分辨率、高精度的多波束资料,这使东海的高精度地形自动分类成为可能。本文选取东海陆坡区及附近海域为研究对象,充分利用已有的高分辨率、高精度多波束资料,分别利用统计分类和神经网络分类方法,实现对东海陆坡区及附近海域的海底地形的大尺度分类。本文的主要研究工作包括:(1)利用多波束数据建立DEM,并从DEM中提取地形信息因子;(2)利用统计方法基于地形因子对研究区进行地形分类;(3)利用人工神经网络方法基于地形因子对研究区进行地形分类;(4)对分类结果进行精度评价和对比分析。研究表明,利用多波束数据对海底地形进行分类,分类精度较高,神经网络方法的分类精度要高于统计分类方法。更多还原

【Abstract】 The Seafloor Terrain Classification (STC) is one method of describing terrain characteristic, and analyzing the complicated extent of the seafloor terrain growth characteristic. It can divide the seafloor characteristic into the different terrain unit. The STC is an important content in the field of the seabed geomorphology, but now it is mainly depended on the descriptive manual classification method, which limits the terrain information reliability. The automatizaton of the STC can greatly improve the precise and intuitionistic representation of the digitization terrain analysis. It also makes the possibility of the quantifiable description to the terrain. This quantifiable and objective description provides the basis for the seabed geomorphology synthesis charting, and offers the important reference value for the submarine geology unit division and the geologic structure recognition. These basis terrain information is important for the marine disaster research and the oil gas resource prospecting. Unfortunately, the relative research activity is very seldom at present.The continent slope in the East China Sea, whose terrain is quite complicated, is an active region involved in many research field. Determining impersonally and accurately the shelf break line and the continental slope toe line in the East China Sea’s research, has very profound scientific significance and the application value. In the East China Sea, the multi-beam sidescan data with high resolution and accuracy have been already accumulated, which makes the highly accurate and automatic terrain classification possible.In this study, a research area is chosen in the continent slope and the continent slope nearby in the East China Sea. And the multi-beam bathymetric data with both high resolution and accuracy are used here. The large scale classification of the study area is carried out by means of both the statistical method and the artificial neural network method, respectively. The main research works include:(1) Establishing DEM (Digital Elevation Model) based on the multi-beam data, and extracting the topographic feature factor from DEM;(2) Carrying out the topographic classification in the research area by means of the statistical method based on the topographic feature factor;(3) Realizing the topographic classification in the research area by means of the artificial neural network method based on the topographic feature factor;(4) Implementing the accuracy evaluation and contrast analysis as a result to the classification.The research indicates that the multi-beam data can give the high precision of the STC. The STC using the artificial neural network method is much more accurate than that from the statistical classification method.更多还原