当前位置: X-MOL 学术Weed Technol. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Flowering leafy spurge (Euphorbia esula) detection using unmanned aerial vehicle imagery in biological control sites: Impacts of flight height, flight time and detection method
Weed Technology ( IF 1.3 ) Pub Date : 2020-01-13 , DOI: 10.1017/wet.2020.8
Xiaohui Yang , Anne M. Smith , Robert S. Bourchier , Kim Hodge , Dustin Ostrander

Leafy spurge, a noxious perennial weed, is a major threat to the prairie ecosystem in North America. Strategic planning to control leafy spurge requires monitoring its spatial distribution and spread. The ability to detect flowering leafy spurge at two biological control sites in southern Saskatchewan, Canada, was investigated using an unmanned aerial vehicle (UAV) system. Three flight missions were conducted on June 30, 2016, during the leafy spurge flowering period. Imagery was acquired at four flight heights and one or two acquisition times, depending on the site. The sites were reflown on June 28, 2017, to evaluate the change in flowering leafy spurge over time. Mixture tuned matched filtering (MTMF) and hue, intensity, and saturation (HIS) threshold analyses were used to determine flowering leafy spurge cover. Flight height of 30 m was optimal; the strongest relationships between UAV and ground estimates of leafy spurge cover (r2 = 0.76 to 0.90; normalized root mean square error [NRMSE] = 0.10 to 0.13) and stem density (r2 = 0.72 to 0.75) were observed. Detection was not significantly affected by the image analysis method (P > 0.05). Flowering leafy spurge cover estimates were similar using HIS (1.9% to 14.8%) and MTMF (2.1% to 10.3%) and agreed with the ground estimates (using HIS: r2 = 0.64 to 0.93, NRMSE = 0.08 to 0.25; using MTMF: r2 = 0.64 to 0.90, NRMSE = 0.10 to 0.27). The reduction in flowering leafy spurge cover between 2016 and 2017 detected using UAV images and HIS (8.1% at site 1 and 2.7% at site 2) was consistent with that based on ground digital photographs (10% at site 1 and 1.8% at site 2). UAV imagery is a useful tool for accurately detecting flowering leafy spurge and could be used for routine monitoring purposes in a biological control program.

中文翻译:

在生物控制点使用无人机图像检测开花的大戟 (Euphorbia esula):飞行高度、飞行时间和检测方法的影响

绿叶大戟是一种有毒的多年生杂草,是北美草原生态系统的主要威胁。控制绿叶大戟的战略规划需要监测其空间分布和传播。使用无人驾驶飞行器 (UAV) 系统研究了在加拿大萨斯喀彻温省南部的两个生物控制点检测开花大戟的能力。2016 年 6 月 30 日,在枝繁叶茂的大戟开花期间,进行了三次飞行任务。图像是在四个飞行高度和一到两个采集时间采集的,具体取决于站点。这些网站于 2017 年 6 月 28 日进行了重新评估,以评估开花的绿叶大戟随时间的变化。混合调谐匹配滤波 (MTMF) 和色调、强度和饱和度 (HIS) 阈值分析用于确定开花的大戟覆盖。30 m 的飞行高度是最佳的;无人机与多叶大戟覆盖的地面估计之间最强的关系(r2= 0.76 至 0.90;归一化均方根误差 [NRMSE] = 0.10 到 0.13)和茎密度(r2= 0.72 至 0.75) 被观察到。图像分析方法对检测的影响不显着(P > 0.05)。使用 HIS(1.9% 到 14.8%)和 MTMF(2.1% 到 10.3%)估计开花的大戟覆盖率相似,并且与地面估计值一致(使用 HIS:r2= 0.64 至 0.93,NRMSE = 0.08 至 0.25;使用 MTMF:r2= 0.64 至 0.90,NRMSE = 0.10 至 0.27)。使用无人机图像和 HIS 检测到的 2016 年至 2017 年间开花的大戟覆盖减少(站点 1 为 8.1%,站点 2 为 2.7%)与基于地面数码照片的结果一致(站点 1 为 10%,站点为 1.8%) 2)。无人机图像是准确检测开花的大戟开花的有用工具,可用于生物控制程序中的常规监测目的。
更新日期:2020-01-13
down
wechat
bug