Tomato spotted wilt virus is a globally prevalent plant disease that threatens thousands of plants in a persistent and propagative way. Early disease identification is predicted to be capable of controlling disease transmission, facilitating management practice, and ensuring associated economic advantages. Hyperspectral imaging, a powerful remote sensing tool has been widely applied in different science fields, especially in the plant science domain. Rich spectral information makes disease detection possible before visible disease symptoms show up. The molecular-level direct detection method can accurately evaluate plant disease levels, but it’s hard to be conducted real-time field tests scientific. Comparatively, machine vision-based indirect detection method is more attractive in practice because of their non-invasive properties and their ability to identify plant diseases through various parameters including color, morphological, and temperature changes. Hyperspectral imaging (HSI) makes use of the plant's interaction with different electromagnetic spectra and forms an image containing the intrinsic information of the leaf biochemical compounds and leaf anatomical structure. Our research will focus on the view of the timeline, and observe how hyperspectral images changes with time before visible symptoms in plants show up. Currently, the experiment only proves TSWV is distinguishable as early as possible. To further improve the early-stage disease detection efficiency, other information like leaf temperature and chlorophyll content is expected to be integrated into the proposed model.
Dr. Md. Monirul Islam
ASRBC, ACI Seed
Krezhova, D., Petrov, N., & Maneva, S. Hyperspectral remote sensing applications for monitoring and stress detection in cultural plants: viral infections in tobacco plants. In Remote Sensing for Agriculture, Ecosystems, and Hydrology XIV (Vol. 8531, p. 85311H). International Society for Optics and Photonics (2012, October).