A non-destructive and precise root monitoring system for hydroponic crops using SWIR hyperspectral imaging

Growing smarter sustainable cities call for a moderate increase in urban agriculture to alleviate burden on traditional agricultural lands. In response, vertical hydroponic farms have emerged as a popular solution, efficiently utilizing limited urban space to produce crops. To ensure consistent, year-round crop quality and maximum yield, continuous monitoring of these farms is crucial. In particular, healthy crop roots are vital for plant growth, as they absorb water and nutrients essential for the growth. Monitoring root dimensions, color, water content, and exudation process provides valuable insights into the overall plant health. However, current root monitoring methods are often contact-based, time-consuming, destructive, subjective, and require sample preparation thereby limiting the potential for future automation. Hence, there is a need for a non-contact and non-invasive approach for root health monitoring based on visual features and root exudate quantification. In this context, this research proposes the development of a non-destructive root monitoring system using a short-wave infrared (SWIR) hyperspectral imager. The proposed method can serve as an excellent system to study the root exudation process and associated root characteristics such as root exudation location and type. The research also proposes the use of a new index termed 'Root Health Index (RHI)' based on the wavelength-specific spectral mapping, and Spectral Angle Mapper (SAM) classification, for evaluating root health. The developed system is demonstrated to enable timely detection of crop salinity stress where a significant reduction in fresh weight (∼62.5 %) and root length (∼21.5 %) was observed.