摘要
利用光谱技术分析了小吉丁虫危害不同等级的塞威氏苹果的反射光谱特征,并用光谱数据针对塞威氏苹果树的虫害进行定量化分析。用新疆维吾尔自治区伊犁哈萨克自治州巩留县野果林实测60条不同虫害等级的塞威氏苹果高光谱数据,分析健康、轻度、中度、重度4个虫害等级光谱反射率及一阶微分光谱特征,建立了红边参数,并在6个检验参数的基础上,构建了虫害等级的检测模型,并用验证组数据对模型进行精度检验。结果表明:(1)健康状态下的塞威氏苹果光谱反射率较受虫害塞威氏苹果光谱反射率高,受害程度越严重,反射率越低。(2)受虫害塞威氏苹果光谱特征表现为“绿峰”红移;“红边位置”蓝移,尤其是受重度虫害的极为明显;近红外反射峰向短波方向移动。(3)红边比值植被指数(RERVI)、红遍斜率(RES)、红边差值植被指数(REDVI)、红边面积(REA)这4个参数与虫害等级呈极显著相关关系,红边归一化植被指数(RENDVI)及红边位置(REP)与虫害等级相关性不强。(4)以RERVI、RES、REDVI、REA这4个参数为自变量构成的多元回归模型的检测精度与准确度均为0.7以上。因此,通过相关参数及模型可以有效检测小吉丁虫害等级。
The spectral reflectance characteristics of Malus sieversii were analyzed by spectral techniques under different levels of damage by Agrilusmali Matsumura , and the spectral data were used to conduct a quantitative analysis on the insect pests of Malus sieversii trees. The 60 Malus sieversii hyperspectral data with different pest damage levels in the wild fruit forest of Gongliu County, Ili Kazakh Autonomous Prefecture, Xinjiang Uygur Autonomous Region were used to analyze the spectral reflectance and first-order derivative spectral characteristics of insect pests that were divided into four levels-healthy, mild, moderate and severe, and create red edge parameters. A detection model of pest damage levels was constructed on the basis of six test parameters, and the accuracy of model was tested by the data of verification group. The results showed that the spectral reflectance of Malus sieversii under healthy condition was higher than that of Malus sieversii damaged by pests, and the more serious the damage was, the lower the reflectivity was. The spectral characteristics of Malus sieversii destroyed by pests were that “green peak” position showed red shift, and “red edge position” showed blue shift, especially for the severity level of pests;the near-infrared reflection peaks moved towards the short waves. The red edge ratio vegetation index ( RERVI ), red edge slope ( RES ), red edgedifference vegetation index ( REDVI ) and red edge area ( REA ) were significantly correlated with the pest damage levels, but the correlation between red edge normalized difference vegetation index ( RENDVI ) and red edge position ( REP ) and pest levels was not strong. The detection precision and accuracy of multiple regression model composed of RERVI, RES, REDVI and REA as independent variables were above 0.7. Therefore, the pest damage levels of Agrilusmali Matsumura can be effectively detected through the establishment of relevant parameters and models.
作者
罗青青
黄铁成
陈蜀江
陈孟禹
贾翔
朱选
来风兵
武红敢
赵文霞
李春蕾
姚艳霞
LUO Qing-qing;HUANG Tie-cheng;CHEN Shu-jiang;CHEN Meng-yu;JIA Xiang;ZHU Xuan;LAI Feng-bing;WU Hong-gan;ZHAO Wen-xia;LI Chun-lei;YAO Yan-xia(School of Geography and Tourism, Xinjiang Normal University, Urumqi 830054, China;Urumqi Institute of Space Remote Sensing Applications, Urumqi 830054, China;Key Laboratory of Precision Forestry,Beijing Forestry University, Beijing 100083,China;School of Foreign Languages,Suzhou University of Science and Technology, Suzhou 215000,China;Monash University, Melbourne 3800, Australia;Reacher Institute of Rescource Information Techniques, Chinese Academy of Forestry, Beijing 100091, China;Key Laboratory of Forest Protection of the State Forestry Administration Research, Institute of Forest Ecology Environment and Protection, Chinese Academy of Forestry, Beijing 100091, China)
出处
《江苏农业学报》
CSCD
北大核心
2019年第4期798-803,共6页
Jiangsu Journal of Agricultural Sciences
基金
国家重点研发计划项目(2016YFC0501503)
关键词
小吉丁虫
虫害等级
塞威氏苹果
反射率光谱
定量化测评
Agrilusmali Matsumura
pest damage level
Malus sieversii
spectral reflectance
quantitative evaluation
