Dataset and model for the paper "Tackling Small Lunar Impact Crater Classification: A Novel Augmented Dataset and Enhanced Deep Learning Framework"

This study proposes a classification framework based on an optimised YOLOv7 model, specifically designed for multi-type small crater identification. We developed the Small Crater Augmented Dataset (SCAD), employing advanced data augmentation strategies to address class imbalance, notably enhancing representation of rare morphologies (flat-bottomed craters, central mounds, and concentric craters). The model integrates focus loss to mitigate class imbalance and adopts efficient intersection-over-union ratio optimisation to enhance boundary detection accuracy.