浙江农业科学 ›› 2026, Vol. 67 ›› Issue (10): 2475-2479.DOI: 10.16178/j.issn.0528-9017.20260256

• 植保技术 • 上一篇    下一篇

基于多光谱遥感与改进YOLOv8的林业有害生物灾害实时监测

徐名川()   

  1. 潮州市潮安区生物多样性保护中心,广东 潮州 521000
  • 收稿日期:2026-04-08 出版日期:2026-10-11 发布日期:2026-10-19
  • 作者简介:徐名川,主要研究方向为林业有害生物防控及生物多样性保护。E-mail:XuMingchuanAA@126.com。

Real time monitoring of forest pest disasters based on multispectral remote sensing and improved YOLOv8

XU Mingchuan()   

  1. Chao’an Biodiversity Conservation Center,Chaozhou City,Chaozhou 521000,Guangdong
  • Received:2026-04-08 Online:2026-10-11 Published:2026-10-19

摘要:

鉴于传统林业有害生物灾害监测方法主要依赖人工巡查与地面调查,存在效率低下、成本高昂、覆盖范围有限等问题,本文提出基于多光谱遥感与改进YOLOv8的林业有害生物灾害实时监测方法。利用无人机搭载多光谱相机获取林业有害生物灾害的多光谱遥感图像,随后进行平滑与变换处理,以此提升图像质量并降低噪声干扰。通过引入注意力机制对常规YOLOv8模型进行改进,并利用改进后的模型对多光谱遥感图像进行分类识别,从而实现林业有害生物灾害等级的实时监测。案例分析结果表明,研究方法可以兼顾较高的林业有害生物灾害实时监测精度与速度,满足大规模林业有害生物灾害的防控需求。

关键词: 多光谱遥感, 改进YOLOv8, 林业有害生物灾害, 灾害监测, 实时监测

Abstract:

Given that traditional forest pest monitoring methods mainly rely on manual inspections and ground surveys,which have problems such as low efficiency,high cost,and limited coverage,this paper proposes a real-time monitoring method for forest pest disasters based on multispectral remote sensing and improved YOLOv8. Multispectral remote sensing images of forest pest disasters are acquired using a drone equipped with a multispectral camera, followed by smoothing and transformation processing to enhance image quality and reduce noise interference. By introducing attention mechanism to improve the conventional YOLOv8 model,and using the improved model to classify and recognize multispectral remote sensing images,real-time monitoring of forest pest disaster levels can be achieved. The case analysis results indicate that the research method can balance high real-time monitoring accuracy and speed of forest pest disasters,meeting the prevention and control needs of large-scale forest pest disasters.

Key words: multispectral remote sensing, improved YOLOv8, forest pest disasters, disaster monitoring, real time monitoring

中图分类号: