Location: Pest Management and Biocontrol ResearchTitle: Identifying optimal wavelengths as disease signatures using hyperspectral sensor and machine learning
|WEI, XING - Virginia Tech|
|AGUILERA, MARCELA - Virginia Tech|
|LANGSTON, DAVID - Virginia Tech|
|LI, SONG - Virginia Tech|
Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/14/2021
Publication Date: 7/19/2021
Citation: Wei, X., Aguilera, M., Langston, D.B., Mehl, H.L., Li, S. 2021. Identifying optimal wavelengths as disease signatures using hyperspectral sensor and machine learning. Remote Sensing. 13. https://doi.org/10.3390/rs13142833.
Interpretive Summary: Plant diseases adversely affect the production of fruit, vegetables, and field crops. The development of cost-effective management practices for disease control relies on accurate and efficient detection and diagnosis of their signs or symptoms. Current scouting methods are mainly based on visual inspection by human raters and are labor-intensive and time-consuming. Recent advances in sensor technologies and data analytic tools enable the development of automated disease detection methods. Hyperspectral sensors measure the spectral reflectance of an object over a wide range of wavelengths, but the large volumes of complex data produced make data analysis challenging. Machine learning utilizes programming algorithms that allow computers to identify trends and patterns from large amounts of data. This study demonstrated the identification of optimal wavelengths to detect a soilborne disease of peanut, stem rot caused by the fungal pathogen Athelia rolfsii, using a handheld hyperspectral sensor and machine learning. Results can be applied to develop methods for automated detection of stem rot in peanut fields using optical sensors. Furthermore, the methodology presented here can be adapted to plant disease detection via hyperspectral sensors in other plant-pathogen systems.
Technical Abstract: Hyperspectral sensors combined with machine learning are increasingly utilized in agricultural crop systems for diverse applications, including plant disease detection. This study was designed to identify the most important wavelengths to discriminate between healthy and diseased peanut (Arachis hypogaea L.) plants infected with Athelia rolfsii, the causal agent of peanut stem rot, using in-situ spectroscopy and machine learning. In greenhouse experiments, daily measurements were conducted to inspect disease symptoms visually and to collect spectral reflectance of peanut leaves on lateral stems of plants mock-inoculated and inoculated with A. rolfsii. Spectrum files were categorized into five classes based on foliar wilting symptoms. Five feature selection methods were compared to select the top 10 ranked wavelengths with and without a custom minimum distance of 20 nm. Recursive feature elimination methods outperformed the chi-square and SelectFromModel methods. Adding the minimum distance of 20 nm into the top selected wavelengths improved classification performance. Wavelengths of 505, 690 to 694, and 884 nm were repeatedly selected by two or more feature selection methods These selected wavelengths can be applied in designing optical sensors for automated stem rot detection in peanut fields. The machine-learning-based methodology can be adapted to identify spectral signatures of disease in other plant-pathogen systems.