摘要
对病虫害进行监测和预警是精准农业的一个重要研究方向,可有效提高粮食产量和质量。本文从使用高光谱遥感技术监测农作物病虫害的原理和技术路线出发,首先阐述了4种光谱特征提取和变换方法,包括原始光谱的导数变换与对数变换、基于光谱位置和面积的特征参数、基于连续统去除的特征参数和植被指数;然后论述了基于统计分析、机器学习和物理模型的3种病虫害监测参数反演方法,建立了光谱反射率和病虫害监测参数间的回归关系;最后分析了将高光谱遥感技术用于农作物病虫害监测的优势和存在问题。
Monitoring and early warning of pests and diseases in crops is an important direction in precision agriculture, which can effectively improve grain yield and quality. From the principle and technical route of monitoring crop pests and diseases with hyperspectral remote sensing, this paper first expounds four spectral feature extraction and transformation methods, including the derivative transformation and logarithmic transformation of original spectra, the characteristic parameters based on spectral position and area ,the feature parameters of continuum removal, and the vegetation indexes. Then this paper discusses three inversion methods, including the statistical analysis, the machine learning and the physical model, by which the regression relationship between spectral reflectance and pest monitoring parameters is established.Finally this paper analyzes the advantages and problems of using hyperspectral remote sensing technology for crop pests and diseases monitoring.
作者
简俊凡
何宏昌
王晓飞
李月
JIAN Junfan;HE Hongchang;WANG Xiaofei;LI Yue(Electronic Engineering College of Heilongjiang University,Harbin 150000,China)
出处
《测绘通报》
CSCD
北大核心
2018年第9期24-28,共5页
Bulletin of Surveying and Mapping
基金
国家重点研发计划(2016YFB0502502)
黑龙江省物联网感知层与传感网络技术创新服务平台
关键词
遥感
高光谱
监测
农作物
病虫害
remote sensing
hyperspectral
monitoring
crop
pests and diseases
