Topographic surface roughness analysis based on image processing of terrestrial planet

We discuss a roughness problems in this paper. Roughness measurements play a fundamental role in terrestrial planet research. The main question discussed in this paper is how to build a more efficient roughness model on the basis of images. The model needs to beat a series of traditional roughness problems on the image. Lunar reconnaissance orbiter camera (LROc) and charged coupled device (CCD) images are our primary data target. Roughness is divided into trend surface roughness and topography roughness with analysis and inversion. The two parts are combined and analyzed. The analysis is based on two structural features: trend reflection and detail details surface. Our team use inversion to remap the surface roughness by CCD image, and then classify the roughness into the trend influence class and the detail influence class. We use frequency analysis tools are to find the discontinuous energy in the frequency domain. Finally, these roughness information are remapped into a visual image results. Our model provide a more accurate method on terrestrial planetary surface roughness problem. We can analyze the difference between the trend surface and the detail roughness by single image. We find that, lunar surface roughness presents more detail with the help of the frequency analysis. Highlands and the mare shows very different roughness decomposition. Our analysis shows better results in the areas in which details surface is not associated with roughness. It is hoped that this study will provide new insights in the surface roughness of terrestrial planets. Especially for CCD images (LROc) rather than details surface data (DEM). In the future, we will focus on calculations and analysis for Mars based on the mode in this paper. For the roughness data, we will start depth learning, classification, pattern recognition and other research in future work.