Using spectral sensing in plant science

Abstract

Digital cameras are widely used tools for plant monitoring in plant science today. Used to track plant growth or even visible symptoms, they are important tools for breeding and plant protection field trials. Nevertheless, its extension to measure the near infrared (NIR) region (700–1000 nm) includes great potential as plants show a higher light reflectance within this spectrum. Various applications have shown its use for disease detection, quantification, virus content estimation, and stress monitoring. As the next step is a comprehensive integration into agricultural routines, this study will show two use-cases with a high technological readiness level. One use-case shows a handheld multispectral sensor, which is used for manual measurements to detect and discriminate different virus types in sugar beet. In contrast, the second use-case shows a transfer to an UAV based disease quantification routine based on spectral imaging for Cercospora leaf spot. In addition, two prototypical workflows are shown for processing non-imaging and imaging spectral data in an agricultural setting. This study shows the state of the art in spectral sensing in the field for the two major sugar beet diseases – virus yellows and Cercospora leaf spot. Furthermore a future perspective for coming technological challenges regarding the integration of AI in sensors or robotic workflows is provided.