用微信扫码二维码

分享至好友和朋友圈

WE ARE COMMITTED TO REPORTING THE LATEST FORESTRY ACADEMIC ACHIEVEMENTS

曾全, 蒲远凤, 肖银波, 等. 基于无人机高光谱的川南疫木林区早期监测研究[J/OL]. 四川林业科技, 2024, 45[2024-04-12]. DOI: 10.12172/202403080001
引用本文: 曾全, 蒲远凤, 肖银波, 等. 基于无人机高光谱的川南疫木林区早期监测研究[J/OL]. 四川林业科技, 2024, 45[2024-04-12]. DOI: 10.12172/202403080001
ZENG Q, PU Y F, XIAO Y B, et al. Early surveillance of pestilence forest area in southern Sichuan based on UAV hyperspectrum[J/OL]. Journal of Sichuan Forestry Science and Technology, 2024, 45[2024-04-12]. DOI: 10.12172/202403080001
Citation: ZENG Q, PU Y F, XIAO Y B, et al. Early surveillance of pestilence forest area in southern Sichuan based on UAV hyperspectrum[J/OL]. Journal of Sichuan Forestry Science and Technology, 2024, 45[2024-04-12]. DOI: 10.12172/202403080001

基于无人机高光谱的川南疫木林区早期监测研究

Early surveillance of pestilence forest area in southern Sichuan based on UAV hyperspectrum

  • 摘要: 为探明马尾松感染松材线虫病早期地理位置及发病率。2021年7月上旬,利用无人机搭载高光谱成像仪采集遥感影像,选用支持向量机进行监督分类,在早期感病反演模型基础上,顺利提取了感病早期的马尾松地理位置及相关信息。结果表明:(1)利用460 nm、525 nm和635 nm的3波段组合真彩色影像进行ROI勾绘,马尾松与其他地被物分离度较高;(2)基于支持向量机的监督分类,顺利获取741株马尾松地理位置及高光谱反射率数据;(3)结合监测模型提取64株疑似感病马尾松,通过随机采样及镜检,马尾松聚类范围感病植株提取准确率86.67%,即马尾松林间发病率7.49%。综上,初步揭示川南地区马尾松林自然状态下松材线虫发病率,有利于今后指导松材线虫病早期精准防治。

     

    Abstract: To investigate the early geographical location and incidence of pine wood nematode disease infected by Masson pine. In early July 2021, remote sensing images were collected by a hyperspectral imager mounted on UAV, and support vector machines were selected for supervised classification. Based on the inversion model of early disease susceptibility, the geographical location and related information of M. pine in the early stage of disease susceptibility were successfully extracted. The results showed that: (1) the combination of true color images at 460 nm, 525 nm and 635 nm was used for ROI mapping, and the separation degree between M. Pine and other ground cover was high; (2) The geographic location and hyperspectral reflectance data of 741 Masson pines were obtained successfully based on the supervised classification of support vector machine; (3) Combined with the monitoring model, 64 suspected infected M. pine were extracted. Through random sampling and microscopic examination, the accuracy rate of infected plants in the cluster range of M. pine was 86.67%, the incidence rate of M. pine in the forest was 7.49%. In conclusion, the incidence of pine wood nematode in the natural state of M. pine forest in southern of Sichuan was preliminatively revealed, which is helpful to guide the early and accurate control of pine wood nematode disease in the future.

     

/

返回文章
返回