Research Progress of Lunar Impact Crater Detection

摘要

Impact craters are the most prominent topographic feature on the lunar surface, which will play a very important role in the construction of lunar bases and lunar surface activities in the future. Due to the large number of meteorite craters, different diameters and shapes, and the superposition of their morphologies, the artificial identification of meteorite craters is not only inaccurate but also inefficient. This paper first summarizes the significance of lunar impact crater research, then analyzes and summarizes the method of crater extraction, research status and the resulting data set, and finally summarizes the difficult problems and directions of future lunar impact crater identification research. Current crater-detection approaches fall into three categories, such as artificial detection, traditional automatic recognition and machine learning dtetection. With the development of computer, the convolutional neural network has become the trend of crater identification. Developing new images or data and training different network models is the key content of future research.