Abstract:

 Moon surface features have great significance in understanding and reconstructing the lunar geological evolution. Linear structures like rilles and ridges are closely related to the internal forced tectonic movement. The extremely rare availability of samples and the difficulty for field works make remote sensing the most important approach for planetary studies. New and advanced lunar probes launched by China, U.S., Japan and India provide nowadays a lot of highquality data, especially in the form of highresolution Digital Terrain Models (DTMs), bringing new opportunities and challenges for feature extraction on the moon. The aim of this study is to recognize and extract lunar features using geomorphometric analysis based on multiscale parameters and multiresolution DTMs. The considered digital datasets include CE-1LAM (Chang’E One, Laser Altimeter) data with resolution of 500 m/pix, LRO-WAC (Lunar Reconnaissance Orbiter, Wide Angle Camera) data with resolution of 100 m/pix, LRO-LOLA (Lunar Reconnaissance Orbiter, Lunar Orbiter Laser Altimeter) data with resolution of 60 m/pix, and LRO-NAC (Lunar Reconnaissance Orbiter, Narrow Angle Camera) data with resolution of 25 m/pix. We introduced landform curvature to recognize the linear structures including Rilles and Ridges. Different window scales and thresholds are considered for feature extraction. Quantitative analyses of the test results in the study area show that the automatic extraction method of linear structures based on landform curvature is effective and feasible; the results can provide important reference for the lunar surface linear structural interpretation and improve the structural interpretation efficiency and accuracy.

Key words:  linear structures, automatic detection, landform curvature, geomorphometric analysis