Lunar Crater Detection Using YOLOv2 Deep Learning

This study presents an approach for detecting lunar craters utilizing the YOLOv2 (You Only Look Once) object detector, a deep learning technique known for its speed and accuracy. The process involves using a comprehensive lunar crater dataset as the input for the YOLOv2 model. The detection process starts with preprocessing the input images to ensure compatibility with the YOLOv2 framework. The model undergoes training with specific options tailored to optimize crater detection, such as adjusting learning rates, batch sizes, and epoch numbers. During training, the dataset is split into training and validation sets to monitor the model’s performance and prevent overfitting. The YOLOv2 model generates detection boxes around the craters, classifying them accurately within these regions. The results demonstrate the efficacy of YOLOv2 in identifying craters with high precision and recall rates, offering a robust tool for lunar surface analysis and aiding in various astronomical research and exploration missions. This approach highlights the potential of advanced deep learning techniques in enhancing our understanding of extraterrestrial geology.

Keywords: Lunar Crater dataset, Pre-Processing, You Only Look Once object detection, Deep learning, Classification.