DBNtech

Automatic detection system of craters on the Lunar surface

This system uses Deep-learning image recognition technology to recognize the lunar ground image objects and automatically detect craters. As deep learning-based object recognition technique learns DEM image with shading relief and a large number of unlabeled craters can be detected, so it contributes to securing high-performance terrain object image processing and automatic statistical technology in space construction.

1) Technology Introduction : Automatic detection system of craters on the Lunar surface
Detection of craters on the lunar surface is one of the most important research areas in aerospace. Traditionally, detecting craters on the lunar surface has been determined by experts’ visual check-up with high-resolution DEM (Digital Elevation Model) images. However, the results of the crater detection varied among experts so that it was difficult to maintain the reliability and consistency of the results. DBNtech conducted a study on automatic crater detection system using Deep Learning-based object detection technology as a consignment project of KICT(KOREA INSTITUTE of CIVIL ENGINEERING and BUILDING TECHNOLOGY). Using various techniques such as "Hill Shade" techniques as a preprocess instead of "DEM images" that are difficult to distinguish by the naked eye, we improved object detection performance of the Deep-learning model. As a result, not only was it possible to detect the large and small craters more precisely, but also a large number of craters that could not be detected with the naked eye could be detected. The result of this study can be also applied in other areas such as construction and transportation infrastructure using Deep-learning based object recognition in addition to space construction, and will contribute to revitalizing extreme environmental construction technologies and creating new markets related to future space exploration.

Start of Content