中文摘要：

利用＂嫦娥一号＂CCD相机获得的遥感影像及其经三线阵数字摄影测量处理后的DEM数据（空间分辨率为500m）进行了月表形貌特征分析.结果表明：月球的平均高程为–742m,最大高程点和最小高程点均位于月球的背面,其中前者位于Engel’gardt撞击坑东缘,后者位于Minkowski撞击坑的次级撞击坑内;月球表面相对平坦,大部分坡度在15°以下,占月球总面积的90%,月球高地的坡度值变化较大,平均坡度大于7°,月海坡度变化较小,多数在3°以下;月表起伏度计算的最佳统计窗口为16km2,大部分区域起伏度在200m以下,而起伏度大于2000m的大起伏山地主要分布在撞击坑周围的环形山区域;基于高程、坡度、起伏度及影像灰度值的归一化处理数据,利用最大似然法进行月海和月球高地的自动提取,其结果与美国地质调查局出版的地质图进行精度评价,获得Kappa系数为0.78,该方法可以较好地进行月球形态特征的提取及分析.

英文摘要：

The satellite Chang’E-1（CE-1） was launched in October 24, 2007 in China, which obtained a large number of scientific data of the global moon, including imageries and laser altimetric data. In this paper, the characteristics of the lunar surface are analyzed and the lunar morphologic information of the lunar mare and highland is automatically extracted based on the imagery acquired by CCD camera and the DEM processed by three-line digital photogrammetric technology. The results show that the average elevation of the global moon is –742 m; the highest and lowest locations stand in the farside of the moon; the highest is at the eastern margin of Engel’gardt crater; the lowest is in the secondary impact crater pit of Minkowski crater. The surface of the moon is very flat; most of its slope is bellow 15°, accounting for 90% of the total area of the global moon. The great variation of slope value mainly distributes in the lunar highlands, with an average slope of greater than 7°. The lunar mares widely distribute in the nearside of moon, with an average slope of below 3°. The optimum statistical unit for lunar relief value is 16 km2. The mare and highland are the basic types of lunar morphology, which can be automatically extracted by maximum likelihood classifier based on the elevation data and slope, relief and imagery data. The values of altitude, slope, relief and grey level of CE-1 imageries are normalized and binarized. A case study of H010 （locating in 0°–18°W, 0°–14°S） shows that the overall accuracy assessment for extracted results of the lunar mare and highland based on the geological data published by USGS in 1971 is better, and the Kappa coefficient is 0.78, and the results of the lunar mare and highland show that the method can be used to extract the characteristics of lunar surface.