基于YOLOv5 的高分辨率遥感图像目标检测算法  

Target Detection Algorithm in High-Resolution Remote Sensing Images Based on YOLOv5

作　　者：丁晋湘 DING Jinxiang(School of Electrical and Information Engineering,Yunnan Minzu University,Kunming 650504,China)

机构地区：[1]云南民族大学电气信息工程学院,云南昆明650504

出　　处：《电视技术》2024年第7期20-23,共4页Video Engineering

摘　　要：随着遥感技术的快速发展,高分辨率遥感图像以其高精度的地表细节信息,为城市规划、环境监测、灾害评估等提供了强有力的支持。但由于目标尺寸多样性和背景复杂性的增加,在高分辨率图像中准确地检测和识别特定目标面临挑战。YOLOv5在实时目标检测领域展现了出色的性能,在处理速度上具有显著优势,在多种复杂场景下保持高准确度。基于YOLOv5算法框架,提出一种面向高分辨率遥感图像的目标检测算法,引入改进多尺度特征融合技术,有效提升模型在复杂背景下对小目标的检测能力。结果表明,所提算法在高分辨率遥感图像目标检测任务中的性能优越,具有实用潜力。With the rapid development of remote sensing technology,high-resolution remote sensing images provide strong support for urban planning,environmental monitoring and disaster assessment with their high-precision surface details.However,due to the increasing size diversity of targets and background complexity,it is challenging to accurately detect and identify specific targets in high-resolution images.YOLOv5 has demonstrated excellent performance in the field of real-time target detection,with significant advantages in processing speed and high accuracy in a variety of complex scenarios.Based on the YOLOv5 algorithm framework,an object detection algorithm for high-resolution remote sensing images is proposed in this paper.Improved multi-scale feature fusion technology is introduced to effectively improve the detection capability of the model for small objects under complex background.The results show that the proposed algorithm has excellent performance in high resolution remote sensing image target detection and has practical potential.

关 键 词：YOLOv5 高分辨率 遥感图像 目标检测 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]