Advances in Plant Disease Detection in Organic Farming Systems

Traditional crop monitoring relies on human observation to determine the presence of disease. As human vision can only sense a narrow segment of electromagnetic spectrum called visible light, spectrophotometers have been used to determine absorbed and reflected light from a surface in a much wider electromagnetic spectrum generating rich hyperspectral data that can support humans with plant disease detection. Furthermore, machine learning techniques to process hyperspectral data can be referred to as the frontier of innovation for non-invasive detection of asymptomatic pathogens as they provide an automatic, real-time and on-field solution for distinguishing healthy and diseased plants with high accuracy. As plants react differently in every component of electromagnetic spectrum, various studies have been conducted to use plants’ hyperspectral data to measure water loss, classify asymptomatic infected samples, assess disease severity, distinguish cultivars of crops, and identify wavelength bands crucial for disease detection. Hyperspectral sensors are also popular in monitoring large fields and orchards with real-time results via unmanned aerial vehicles (UAV).