【作者】 罗柏文；

【导师】 卜英勇；

【作者基本信息】 中南大学 ， 机械设计及理论， 2008， 博士

【摘要】 深海富钴结壳是一种生长在水深500～3000m的平顶海山、海台顶部或坡上的壳状物,厚度一般为2～6cm,在水深800～1500m的高集区,壳厚可达10～20cm。它除含锰、镍、铁、铜等金属外,富含钴、铂等稀贵金属,是一种重要的战略矿产资源。我国在太平洋海底圈定了一块7.5万平方公里的“战略金属资源基地”。由此研究钴结壳的采集技术与装备具有重大的经济价值和战略意义。成功采集钴结壳的关键技术之一是有效控制下层岩石的贫化。在钴结壳矿层厚度和微地形表面特征已知情况下可获得最佳采集深度。国家自然科学基金项目“深海钴结壳微地形监测技术与最佳采集深度建模研究”(50474052)的主要研究内容是寻求适合钴结壳采集的海底微地形探测技术与方法;研究的目的是为采矿车的最佳采集深度计算提供实时的规则格网DEM数据。论文在此课题背景下,从探测方案选取到微地形DEM建立始终围绕着实时性和高精度两大基本要求展开研究工作。旨在为开发出真正能用于海底微地形探测的装置提供理论指导和实验依据。作者首先从水下微地形探测方案着手,查阅大量有关声、光、电的水下地形探测技术资料,并结合钴结壳特殊赋存环境和具体开采方案,提出一种摆动式单波束探测方法。随后,研制出实验用的摆动式单波束探测装置。利用该装置进行一系列的探测实验,为后面的高程数据预处理及DEM建立提供必要的验证数据。其次是实现高程数据预处理。摆动式单波束探测受掠射角、底质特性等因素的影响,不可避免地产生高程异常值。高程异常值是导致“伪”地形的首要因素,加上背向超声波的地形坡面无法被探测以及必须建立规则格网DEM等原因,有必要对高程数据进行预处理。采取的主要措施有:1)通过一种改进的数据加窗法剔除高程异常值;2)采用改进的中值滤波算法实施高程数据滤波;3)通过支持向量回归模型确定横向(X方向)测线高程值;4)利用坡度信息对高程值进行检测与修正。经过预处理后的高程数据简化了微地形规则格网DEM的内插问题。X方向测线上可由支持向量机回归预测任意间隔的高程值,只考虑沿Y方向实施高程值内插,而待定点都处在沿Y方向高程值的连线上,采用线性内插方法可建立分辨率为25mm的规则格网DEM。并且通过几何量计算和逼真性评价检验这种方式建立的微地形DEM达到了工程中要求的精度等级。最后,为了增强DEM的地形表达效果,作者在AVS/Express软件环境下利用VC++编程语言创建一个VMTerrain模块,通过该模块实现微地形可视化。更多还原

【Abstract】 Deep-ocean cobalt-rich crusts (CRC) is the shell covering top-flat mountains, the top or the slop of sea plateau, under the sea ranging from 500 to 3000 meters deep. Generally, the crust is 2～6 centimeters thick, and it can be 10～20 centimeters thick in high enriched areas which is 800～1500 meters deep under the water. CRC contains manganese, nickel, iron, copper and other common metals. Especially, it is rich in rare and precious metals such as cobalt, platinum and so on. That makes it one of important strategic mineral resources. China has designated an area of 75 thousand square kilometers on the Pacific seafloor as the strategic metal resource base. Thus, the research on the exploiting techniques and equipments is of great economic value and strategic importance.To efficiently control the ore dilution is one of the key techniques to successfully exploit cobalt-rich crusts. When the seam thickness of cobalt-rich crusts and the surface feature of micro-terrain are known, the best cutting depth can be acquired.The main research of the project "Research on detection technology of deep-ocean cobalt-rich crusts micro-terrain and best cutting-depth controlling model"(NO.50474052), supported by the National Natural Science Foundation of China, is to find methods and techniques of detecting the micro-terrain of seafloor which is suitable to exploit CRC. Under the background of the project, the research of this dissertation, from the selection of detecting plan to the construction of micro-terrain DEM, consistently revolved around the two fundamental requirements of good detecting real-time property and high detecting precision in order to provide theoretical guide and experimental basis to construct practical equipment to detect seafloor micro- terrain.The author started with the underwater detecting plan of micro-terrain and put forward a detecting method with a pendulum single beam after consulting a large quantity of acoustic, optical and electronic technique materials on detecting underwater micro-terrain and considering the specific surrounding and methods of exploiting CRC. Then the pendulum single beam bathymeter was constructed. A series of detecting experiments were performed using this equipment. These experiments provided necessary testifying data for the subsequent pretreatment of elevation data and DEM construction.Secondly, the pretreatment of elevation data was realized. Abnormal elevation data is the chief factor leading to produce pseudo-terrain. Additionally, the back side of the slope is unable to be probed by ultrasound, regular grid DEM model must be constructed and other constraints exist. It is necessary to pre-treat elevation data, including rejecting abnormal elevation data, forecasting and revising elevation data and so forth. In terms with these, the author obtained many achievements and inventions as follows:1. Am improved method of data window was used to eliminate abnormal elevation values when the pendulum single beam bathymeter was detecting.2. The filtering of elevation data was conducted by improved median filtering algorithm.3. Support vector machine was used to forecast elevation values in X-axis direction detecting trajectory.4. Forecasted elevation was checked and rectified in the light of slope information.The pretreated elevation data simplified the interpolation of regular grid DEM model of micro-terrain. The elevation of arbitrary interval in X-axis direction was able to be forecasted by way of support vector machine, only the interpolation of elevation in Y-axis direction was under consideration. However, candidate points settled on the line of forecasted elevation in Y-axis direction. The regular grid DEM model of resolution of 25mm can be constructed by way of linear interpolation. The author testified by geometric calculation and verisimilitude assessment that the micro-terrain constructed in above method was qualified to the precision level as practical engineering required.Finally, the author enclosed the VMTerrain visualizing module using VC++ in the AVS/Express software environment, and visualized micro-terrain DEM through this module.更多还原