Lunar Crater Detection based YoloV5 using CCD Data

Abstract

This paper proofs that one-stage object detection neural network of YoloV5 also performs well on lunar crater detection. The experiment is based on CCD data provided by LROC camera carried by NASA’s Lunar Reconnaissance Orbiter (LRO), and gets a detection result of nearly 75% on test set without any preprocessing operation on the original image.