The improved model of YOLOV8 integrating LSKNet
DOI:
https://doi.org/10.63313/JCSFT.9007Keywords:
Target detection, Satellite remote sensing, Yolov8, LSKNet backbone networkAbstract
Due to the characteristics of satellite remote sensing targets such as small size, dense distribution, variable angles, and complex backgrounds, traditional detec-tion methods are prone to false detections and missed detections. To improve the detection performance of the model, an improved Yolov8 model for remote sensing small target detection is proposed. LSKNet is adopted as the backbone network, which dynamically adjusts its large spatial receptive field to better model the ranging scenarios of objects in remote sensing scenes. The verifica-tion through the DOTA-v1.0 dataset shows that the average precision (mAP) value of the improved model reaches 60.2%, which is 2.9 percentage points higher than that of the traditional model. The improvement measures can sig-nificantly enhance the accuracy and efficiency of detection, providing a more reliable and efficient solution for the field of satellite remote sensing.
References
[1] CHEN F K, Li S X. Improved Yolov5 for Target Detection in Unmanned Aerial Vehicle [J]. Journal of Computer Engineering & Applications, 2023, 59(18).
[2] LIU R J, He Z M. Rapid Target Detection Method for Satellite Remote Sensing Images Based on YOLOv8 [J]. Space Control Technology and Application, 2023, 49(5): 89-97.
[3] LIU J M. Sliding Window Target Detection in Satellite Remote Sensing Images Based on YOLOv5 [J]. Ship Electronic Engineering, 2023, 43(1): 41-46.
[4] Li Z D, HOU X H, WANG L Y, et al. Target Detection in Aerial Remote Sensing Images Based on Dual Attention YOLO [C]// Proceedings of the 17th National Conference on Signal and Intelligent Information Processing and Applications. School of Information, North China University of Technology; Beijing Beifang Zhitu Information Technology Co., Ltd., 2023: 5. DOI: 10.26914/c.cnkihy.2023.054793.
[5] Li Y, Hou Q, Zheng Z, et al. Large Selective Kernel Network for Remote Sensing Object De-tection. arXiv 2023[J]. arXiv preprint arXiv:2303.09030.
[6] Liu Z, Lin Y, Cao Y, et al. Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2021: 10012-10022.
[7] Yang Q, Meng H, Gao Y, et al. A Real-Time Object Detection Method for Underwater Com-plex Environments Based on FasterNet-YOLOv7[J]. Journal of Real-Time Image Processing, 2024, 21(1): 8.
[8] Liu X, Peng H, Zheng N, et al. Efficientvit: Memory Efficient Vision Transformer with Cas-caded Group Attention[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023: 14420-14430.
[9] Zhu L, Wang X, Ke Z, et al. Biformer: Vision Transformer with Bi-Level Routing Atten-tion[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recog-nition. 2023: 10323-10333.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 by author(s) and Erytis Publishing Limited.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
