Automated Lunar Crater Detection using YOLOv8 on Chandrayaan-2 Imagery

Craters are one of the most noticeable structures on planetary surfaces, which are utilized for spacecraft navigation, hazard identification, and age calculation. A number of factors make crater detection a difficult job, including complicated crater characteristics, variable sizes and forms of the craters, planetary data types, and data resolution. An innovative method for identifying and examining craters on the lunar surface using the remote sensing images from Chandrayaan-2 and employing deep learning techniques is proposed in this research. By making use of the extensive dataset from Chandrayaan-2, the proposed approach, YOLOv8-CCNet, uses convolutional neural networks (CNNs) and YOLOv8 to automatically detect crater features with great accuracy and efficiency. The proposed approach of using modified YOLOv8-CCNET showed an accuracy of 90% and IoU of 0.75. By combining remote sensing data processing with deep learning, the study aims to improve the precision of crater detection and characterization. This analysis helps classify different geological areas on the Moon. The techniques developed in this research not only increase the understanding of the Moon but could also be applied to studying other planets, contributing significantly to the field of planetary science.