【作者】 杨奇；

【导师】 车明；

【作者基本信息】 天津大学 ， 模式识别与智能系统， 2017， 硕士

【摘要】 近年来,随着人们对三维显示的需求越来越大,一种基于3D实体沙盘的全新的显示方式已经成为当前研究的重点。本文中的三维数字沙盘系统将卫星航拍的遥感数据与动态实体沙盘结合起来,通过技术手段将遥感数据从平面搬迁到真实的空间中来,提供了一种全新的展示模式。本文实现了三维数字沙盘系统嵌入式控制器及控制软件部分。其中,为了控制机械装置模型,设计嵌入式控制器的软件和硬件,并针对机械装置模型中行程触点的不稳定所导致的中断误差,提出了从软件层次设计中断请求信号过滤层,解决硬件中断误差的方案。另外,针对使用现场的突发性断电问题,设计防断电数据保护方案以及一套可靠的恢复流程,为装置的恢复过程提供正确的数据。为实现对遥感信息的处理与传递,本文基于GDAL库,设计上位机控制软件的功能。控制软件通过对遥感信息的处理和展示,并提供用户选择工具,取得用户感兴趣的区域;接着,利用GDAL库,完成对遥感数据的抽样、投影以及转换,最终获得符合沙盘终端格式的目标数据,并通过串口传输给终端。当终端的标准单元发生故障时,我们称之为故障点,对故障点的聚类分析有助于找出故障根源。本文中利用限定距离的强约束关系,对传统的模糊等价聚类算法进行改进,实现了两种改进后的方法:利用负反馈迭代微调?值法和根据限定距离约束聚类方法。两种改进方法能够使得聚类结果满足约束条件,提高聚类的准确性。实验检测表明,终端的误差弥补方案的使用,使得机械装置运行的出错概率由原始的47.5%降为0.5%,并且当系统发生意外断电后,沙盘系统的终端能够被正确地恢复。同时,实验检测发现使用嵌入式控制器托举长方体立柱升降相同的高度,所需步距角的差值和总步距角的比例为0.47%,由此表明机械装置模型的移动误差符合工业误差标准。由实验结果分析可知,系统终端对机械装置的控制具有很高的稳定性和可靠性。更多还原

【Abstract】 For the past few years,a completely new method which is based on entity sandbox for 3D display has become the main emphasis of our research.The system of 3D digital sandbox introduced in this thesis combined the remote sensing data and the digital entity sandbox,and moved the remote sensing data from plane to real space,providing a brand-new method to display information.The 3D digital sandbox system included two parts.One is embedded control board in terminal,and the other is desktop control App in PC.In order to control the mechanical device,software and hardware were designed and complemented.Among them,since the errors of hardware of mechanical model lead to abnormality of interruption,a set of scheme that modifying the hardware error in the point view of software and filtering interrupt request signals was proposed.The data protection scheme provided correct data for module recovery,when meeting power failure.Beyond that,in order to handle and transmit the remote sense data,desktop App based on GDAL library and serial port was designed to handle and display data,offer pick tool for users and achieve the picked areas.The target data satisfied data style of terminal were sampled,projected and transformed by GDAL,and would transmit to all terminals.It is called fault point,when standard unit of the terminal has failure.Analysis for fault points with cluster algorithm can help to identify the main cause of the terminal failure.According to the strong distance constraint,the traditional fuzzy clustering algorithm is improved to two kinds of methods: adjustment method based on iteration with feedback and clustering method based on the distance strong constraint.The results show that the using of the terminal error modify scheme reduced the error probability of the mechanical device from the 47.5% to 0.5%.The terminal of system would be recovered correctly when it powered off.It also shows that mechanical device could decrease step angles error to 0.47% that conforms industrial error criteria.更多还原