项目名称: 基于立体匹配和明暗恢复形状集成的嫦娥影像月球精细制图
项目编号: No.41201480
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 地理学
项目作者: 彭嫚
作者单位: 中国科学院遥感与数字地球研究所
项目金额: 25万元
中文摘要: 嫦娥计划获取了覆盖全月范围的轨道器数据,根据这些数据进行精细制图对嫦娥二期工程的着陆巡视探测和相关科学目标研究具有重要意义。由于月表纹理贫乏,目前基于摄影测量的传统制图方法难以满足精细制图的需求,影响了该数据的进一步实际应用。针对上述问题,本项目探讨基于立体匹配和明暗恢复形状算法集成的嫦娥数据的精细制图。在分析明暗恢复形状算法理论基础的前提下,使用开发的自适应马尔科夫算法利用嫦娥立体数据生成初始的DEM,将研究区域进行常量反照率和变化反照率区域的划分。对现有利用多幅影像提取地形的明暗恢复形状算法进行改进,最终提出一种结合线阵立体几何约束条件的明暗恢复形状制图方法。通过本项研究,将实现嫦娥数据的像素级精细制图,研究成果将进一步推动和促进嫦娥影像和其他月球探测轨道器数据的实际应用。
中文关键词: 行星摄影测量;影像匹配;数字高程模型;嫦娥影像;明暗恢复形状
英文摘要: China's Chang'E project has obtained orbital data covering the entire lunar surface. Precision lunar terrain mapping is critical for the followup landing and rover exploration and scientific research. As lunar surface is of poor texture, the traditional photogrammetric mapping method can't provide precise mapping product effectively. This project proposes an integration of photogrammetric method and shape from shading (SFS) method for precise lunar mapping. Based on the SFS principle and initial DEM derived from stereo images, a novel method is developed to separate the surface under into regions with constant albedo and variable albedos.Then geometrical constrained SFS is developed by taking line sensor geometry into consideration. Accuracy evaluation of the proposed method is performed to using laser altemeter data as reference. Terrain data from international lunar missions will be used for cross validation. We expect to achieve a pixel level 3D reconstruction of lunar terrain. The research results will be valuable for precise lunar mapping from Chang'E orbital imagery and other lunar orbital data.
英文关键词: planetary photogrammetry;image matching;DEM;Chang'e image;shape from shading