Advanced Crater and Boulder Detection in Lunar Exploration with CNN and YOLO

Advanced Crater and Boulder Detection in Lunar Exploration with CNN and YOLO

Prof. D.A.Gaikwad1, Vedant Shirpurkar2, Navlesh Todkar3,  Zeel Sutar4  ,Vaibhavi Salunkhe5, Zeel Sutar6

Prof. Dipti Gaikawad , Department of Information Technology, JSCOE

Vedant Shirpurkar, Department of Information Technology, JSCOE

Navlesh Todkar, Department of Information Technology, JSCOE

Zeel Sutar, Department of Information Technology, JSCOE

Vaibhavi Salunke, Department of Information Technology, JSCOE

Vijay Madane, Department of Information Technology, JSCOE

Abstract
Lunar craters play a key role in understanding planetary surfaces, including age estimation, hazard detection, and navigation. However, their detection is challenging due to variations in size, shape, and data resolution. This study introduces YOLOv8, a cutting-edge deep learning model, to automate crater detection with high speed and precision. Trained on high-resolution lunar images, YOLOv8 efficiently identifies craters by predicting bounding boxes and class probabilities in a single pass. Additionally, we review over 140 crater detection methods, categorising them into segmentation-based, object detection-based, and classification-based approaches. Our findings highlight YOLOv8's potential to advance automated crater detection and provide insights for future exploration and research.

Key Words:

Lunar craters, Boulder detection, Orbiter High-Resolution Camera (OHRC), Artificial Intelligence (AI), Machine Learning (ML), Deep Learning, Image Processing, Selenographic Positions, Automated Detection, Lunar Surface Analysis.