Scholarlink

In recent years, the subject of detection of Lunar impact craters based on deep learning methods has been widely studied with the aim of providing efficient and effective methods for the automatic detection of Lunar impact craters. In this paper, we used the new version in the YOLO series models, YOLO v9, to apply them to the detection of Lunar impact craters based on the CCD and DEM data provided by NASA to obtain accurate detection results. The performance results indicate that YOLOv9-c achieves a precision of 73.42% on CCD data, while YOLOv9-e records a slightly higher precision of 75.27%. On DEM data, their precision is 69.43% and 70.31% respectively. Additionally, the newly developed lightweight networks, GELAN-c and GELAN-e, were tested on the same CCD and DEM datasets. GELAN-c reached a precision of 74.84% on CCD data and 70.82% on DEM data. GELAN-e showed a precision of 75.81% on CCD data and 70.87% on DEM data. The results have proved that the new version of YOLO, YOLO v9, performed effectively and excellently in the task of the Lunar impact crater.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Learning Techniques for Lunar Impact Crater Identification Based on CCD and DEM Data

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Siyi Chen,
  • Cheng Li,
  • Yueqi Ma,
  • Jiarui Liang,
  • Jionghao Zhu,
  • Xiaolin Tian

摘要

In recent years, the subject of detection of Lunar impact craters based on deep learning methods has been widely studied with the aim of providing efficient and effective methods for the automatic detection of Lunar impact craters. In this paper, we used the new version in the YOLO series models, YOLO v9, to apply them to the detection of Lunar impact craters based on the CCD and DEM data provided by NASA to obtain accurate detection results. The performance results indicate that YOLOv9-c achieves a precision of 73.42% on CCD data, while YOLOv9-e records a slightly higher precision of 75.27%. On DEM data, their precision is 69.43% and 70.31% respectively. Additionally, the newly developed lightweight networks, GELAN-c and GELAN-e, were tested on the same CCD and DEM datasets. GELAN-c reached a precision of 74.84% on CCD data and 70.82% on DEM data. GELAN-e showed a precision of 75.81% on CCD data and 70.87% on DEM data. The results have proved that the new version of YOLO, YOLO v9, performed effectively and excellently in the task of the Lunar impact crater.