节点文献

无控制DEM匹配与差异探测研究

Research on DEM Matching and Deformation Detection Without Control Points

【作者】 冯义从

【导师】 岑敏仪;

【作者基本信息】 西南交通大学 , 地图制图学与地理信息工程, 2005, 硕士

【摘要】 在计算机视觉、模式识别、工业产品质量控制、医学影像、土壤侵蚀、泥石流灾害监测以及军事应用等领域,常需对两个相同或相近的三维表面进行匹配和在此基础之上的差异探测研究。由于有时布设控制点的困难,无控制三维表面匹配和差异探测技术得到广泛应用,本文主要研究基于规则格网DEM的无控制匹配和差异探测问题。 本文首先对3D自由表面匹配技术从表达方法、转换方式、重构策略和匹配算法等几个方面进行了研究和分析,然后以规则格网DEM为匹配研究对象,采用LZD算法为匹配方案,进行无控制DEM匹配研究和实现。通过分析匹配算法原理和转换参数相对关系,提出了增强LZD算法拉入范围的坡度平滑技术,改善了LZD算法的拉入范围,增强了匹配自动化程度;并针对DEM数据趋于海量化的趋势,结合规则格网DEM易于多尺度表达的特点,提出了基于LOD技术的海量数据DEM匹配解决建议。 在实现无控制DEM匹配基础上,针对匹配应用中表面差异探测需要,应用稳健估计理论进行差异探测研究和试验。在泥石流灾害引起的地表变化探测应用中,提出顾及地形特征的F-LZD+M匹配算法,并通过实验比较分析已有差异探测算法LZD+M和LZD+LMS,与新算法F-LZD+M的效果。仿真试验证明F-LZD+M算法的匹配精度和计算效率比LZD+M估计和LZD+LMS算法优越。最后,在普歪沟流域应用F-LZD+M算法,成功地探测到泥石流灾害引起的地表变化量。实例试验表明F-LZD+M算法在无控制表面匹配探测地表变化中有实用价值。

【Abstract】 The research on 3D free-form surface matching and deformation detecting is vitally essential in the fields of computer vision, pattern recognition, quality control of industry products, medical imaging, soil erosion, debris flow disaster monitoring, and military application. The 3D surface matching and deformation detecting without control points is widely used because of the difficulty to establish control points sometimes. This paper is focused on 3D surface, based on Grid DEM, matching and deformation detecting without control points.This paper firstly researched on existing techniques of the representation, transformation wise, registration strategy and matching algorithm of 3D free-form surface, then makes in-depth research on Grid DEM matching without control points based on LZD algorithm. Through analyzing the algorithm principle and the transformation parameters relationship, the author puts forward slope-smoothing technique to strengthen and improve the pull-in range of LZD and the matching automatization degree. Because Grid DEM can be expressed in a multi-scale way, some propositions are given to solve mass data DEM matching problem based on LOD technique.After the realization of the 3D surface matching without control points, because the requirement of change detecting in surface matching, the research and test on deformation detection are carried out based on robust estimator. In the application of detecting the earth’s surface change resulted from debris flow, a new algorithm, F-LZD+M, in which terrain feature is taken into consideration, is put forward in this paper. The advantages and disadvantages of the new algorithm for detecting surface differences and two existing algorithms, LZD+M, LZD+LMS, are analyzed compared by the experiments. The simulation experiments prove that F-LZD+M is superior to LZD+M and LZD+LMS in the perspective of matching accuracy and efficiency. And finally, the magnitude of earth’s surface change resulted from debris

【关键词】 DEM匹配稳健估计地形特征
【Key words】 DEMMatchingRobust EstimationTerrain Feature
  • 【分类号】P208
  • 【被引频次】2
  • 【下载频次】202
节点文献中: 

本文链接的文献网络图示:

本文的引文网络