The Chang’E-1 orbiter, China’s first lunar probe, obtained a large amount of lunar data, especially the global image of the moon. The distribution characteristics of the impact craters can provide a large amount of information on impact history and the lunar evolution process. So it’s significant to research effective and feasible methods of crater automatic extraction.

WAN Cong, CHENG Weiming and their research team, Institute of Geographic Sciences and Natural Resources Research (IGSNRR), have attempted to compare three methods of automatic extraction of lunar impact craters based on the digital elevation model (DEM) data produced from the Chang’E-1 CCD camera in two research areas. The three methods are direct extraction from flooded DEM data, object-oriented extraction from DEM data by using ENVI ZOOM function and object-oriented extraction from flooded DEM data, which is a novel mothod. After the process of accuracy assessment, extracted degree computation, cumulative frequency analysis, shape and age analysis, the researchers found that the novel method yields better accuracy than the other two methods. The number of impact craters extracted by the three methods in the Imbrian period is the largest and is of various types; as to their age earlier than Imbrain, it is difficult to extract because they could have been destroyed.

The research has been published in Science in China Serics G. (Wan C, Cheng W M, Zhou Z P, et al. Automatic extraction of lunar impact craters from Chang’E-1 satellite photographs. Sci China-Phys Mech Astron, 2012, 55: 162？169, doi: 10.1007/s11433-011-4584-6).

The extracted results of lunar impact craters from three different methods. (a) Flooded; (b) Object-Oriented; (c) Flooded Object-Oriented.(Imaged by Cheng Weiming)