Development of a vision system for safe and high-precision soft landing on the Moon

Modern stage of space exploration is tightly related with the Moon research, with the deployment of long-term research stations on its surface. One of the critical elements here is a system for providing a highly accurate and reliable soft landing. The task of soft landing was solved many years ago, however, its accuracy is insufficient for modern lunar projects. Modern computer vision systems allow getting high-quality surface images at all stages of landing, and processing them in real time. The use of technical vision is beneficial due to the low weight, dimensions and energy consumption. The following tasks are assigned to the vision system: self-position determination of the landing unit and selection of a suitable place for a safe landing. Self-position determination by the video is necessary for INS data correction. Two methods are considered: detection and comparison of craters, and GHT. Both methods use the vector map and don't require actual high-precision orthoplan. The safe site choosing method uses the found position for binding it to the desired and undesirable sites map. After that a precise choice follows, which takes into account the surface angle, smoothness, presence of extensive shadow areas. The algorithm is able to select landing sites close to those manually selected by the operator. The obtained algorithms can work in real time with the influence of interfering factors (noise, changes in the Sun position, changing of camera angle), and can be used to solve the problem. (C) 2021 The Authors. Published by Elsevier B.V.