Feasibility assessment of multi-spectral satellite sensors in monitoring and discriminating wheat diseases and insects

详细信息    查看全文

• 作者：Lin Yuan^a ; ^b ; Haibo Zhang^a ; Yuntao Zhang^a ; Chen Xing^a ; Zhiyan Bao^a ; ^{yuanl_rs@163.com}
• 关键词：Winter wheat ; Diseases and insects ; High-resolution satellite ; Multi-spectral sensor ; Discrimination analysis
• 刊名：Optik - International Journal for Light and Electron Optics
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：131
• 期：Complete
• 页码：598-608
• 全文大小：1348 K
• 卷排序：131

文摘

Monitoring and differentiating multiple crop diseases and insects over vast areas are of practical importance in guiding differential treatment. Despite the hyperspectral remote sensing had shown effectiveness in discriminating diseases and insects, the scarcity of the hyperspectral data limited its application. Recently, a number of high resolution multi-spectral satellite sensors became available around the world, which provided routine observations and thus offered an important chance for remotely sensed monitoring of crop diseases and insects. To assess the suitability of channel settings of these broad-band satellite sensors in monitoring and discriminating crop diseases and insects, a feasibility analysis was conducted relying on hyperspectral spectral experiment and a simulation procedure based on channels’ relative spectral response (RSR) function. Taking three typical diseases and insect of winter wheat as an example, this study analyzed the suitability of channels’ reflectance and some classic vegetation indexes (VIs) of seven high-resolution satellite sensors. For each channel/VI of all sensors, the capability in differentiating healthy and infected samples was assessed by an independent t-test, whereas the capability in discriminating diseases and insect was evaluated by an analysis of variance (ANOVA). The results showed that the response of the channels/VIs to diseases and insect reflected the corresponding spectral changes and were highly consistent among sensors. Based on the selected channels and VIs, a discrimination model was established. The model produced a satisfactory overall accuracy of 0.74, which suggested the feasibility of using high resolution multi-spectral satellite sensors in monitoring and discriminating crop diseases and insects.