Improved YOLO11 for Moon Crater Detection

Impact craters are one of the most prominent features on planetary surfaces. Detection and counting of impact craters can be used for estimating the age of planetary surfaces, selecting landing sites, and spacecraft navigation. Compared with traditional target detection, craters have a dense distribution of targets and a large number of small targets. The lower accuracy in detecting small craters has led to a decrease in the overall detection accuracy of craters. Therefore, this paper builds a STE-YOLO based on YOLO11 and SPDConv. A dataset comprising 4,760 images was generated from the Lunar LRO LROC WAC global mosaic. A comparison test was conducted, and the results showed that the precision of STE-YOLO, when based on these images, was higher than that of YOLO11, from 76.4% to 78.3%, the recall from 72.9% to 74.3%