具有抗差特点嫦娥一号异源数据地形滤波与插值方法研究

发布时间：2018-01-20 09:01

本文关键词： CCD立体相机激光高度计LAM 数据滤波克里金插值法异源数据加权插值融合　出处：《中国地质大学（北京）》2013年硕士论文　论文类型：学位论文

【摘要】：月球是距离地球最近的星球，仅有38.4万公里，因此成为人类深空探测和认识宇宙的第一目标，是开展深空探测的理想的中转站，在未来的能源和资源利用与开发方面起着不可估量的作用。月壤中含有丰富的气体资源、水资源和矿产资源，是未来人类解决能源危机的重要途径，因此了解月球的地质构造和月球的三维地貌具有重要的意义，生成更精确、更具体的三维空间形态也显得尤其重要。本文研究的重点就是通过数据的加权融合生成更为精细、准确的月球三维空间形态，了解更详细的月球三维特征。本文中选取了嫦娥一号CCD立体相机的正视影像图制作的1:250万全月分幅数字高程模型数据和激光高度计LAM获取的高程数据作为两种异源数据，选取了包括虹湾地区在内的四个区域作为研究对象，通过移动曲面拟合法进行了数据滤波，剔除一些粗差点，再运用克里金插值的方法，利用变差函数计算权值的特点，结合C++编程和matlab程序对两种异源数据根据各自的精度分别加权实现异源数据的加权插值融合，生成了更精细、更具体的月球地区的数字高程模型（DEM）。本文中对异源数据加权插值融合生成的DEM模型也进行了误差分析，将三线阵CCD数据的高程值加上随机误差，利用加上随机误差的CCD数据和原来的激光高度计LAM数据作为两套异源数据，，根据精度加权插值融合，生成DEM模型。再将三线阵CCD数据单独插值生成DEM模型，最后利用上述得到的DEM的高程值相减得到高程差值。分析生成的高程差值和加上的随机误差统计分布直方图和统计特征值，比较统计特征值最大值最小值和中误差，说明本论文中在普通克里金插值基础上加权插值融合方法的可行性和有效性。
[Abstract]:The moon is the closest planet to Earth, only 384,000 kilometers away, so it has become the first target of deep space exploration and understanding of the universe, and is the ideal transit station for deep space exploration. The lunar soil is rich in gas resources, water resources and mineral resources, which is an important way for human to solve the energy crisis in the future. Therefore, it is of great significance to understand the geological structure of the moon and the three-dimensional geomorphology of the moon. The emphasis of this paper is to produce a more precise and accurate three-dimensional space shape of the moon by weighted fusion of data, and to understand more detailed three-dimensional features of the moon. In this paper, the digital elevation model data of 1: 2. 5 million month full amplitude division and the elevation data obtained by LAM of laser altimeter are selected as two kinds of heterogenous data. Four regions, including Hongwan area, are selected as the research object. The data filtering is carried out by moving surface fitting method, and some rough handicaps are eliminated, and then the Kriging interpolation method is used. Using variation function to calculate the weight value, combining C programming and matlab program to realize the weighted interpolation fusion of the two kinds of heterogenous data according to their respective accuracy respectively, and to produce more fine. More specific digital elevation models for the Moon region. In this paper, the DEM model generated by weighted interpolation fusion of heterogeneous data is also analyzed, and the height of three-line array CCD data is added with random error. The CCD data with random error and the original LAM data of laser altimeter are used as two sets of heterogenous data, which are fused according to precision weighted interpolation. The DEM model is generated and the three-line array CCD data is interpolated separately to generate the DEM model. Finally, the height difference is obtained by subtracting the height value of DEM, and the generated height difference and the statistical distribution histogram and statistical eigenvalue of random error are analyzed. By comparing the statistical eigenvalue maximum, minimum value and median error, the feasibility and effectiveness of weighted interpolation fusion method based on ordinary Kriging interpolation in this paper are demonstrated.
【学位授予单位】：中国地质大学（北京）
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P184.6

【参考文献】