Prototype of chili pathogen early detection system by using multispectral NIR/NUV

Chili plants ( Capsicum annuum L. ) are a high-value horticultural commodity but are very susceptible to disease. Therefore, early detection of chili disease is essential to minimize the potential loss in chili farming. This research aims to develop a prototype for early detection of chili diseases before they become apparent to the human eye. In response to pathogens, chili plants produce substances that actively absorb and reflect ultraviolet light, while near-infrared images can reveal leaf cell structure damage. By considering these plant defense systems, the prototype system, developed with a closed growth chamber, focuses on capturing NIR and NUV images to detect plant diseases. It uses light reflectance in the near-ultraviolet and near-infrared spectrum as input for detecting diseases, coupled with image and pattern analysis for plants affected by viruses and fungi. The primary method of image analysis was texture analysis of NIR and NUV images, specifically image entropy analysis. The system was tested on plants with virus infection (Gemini virus), fungal infection (anthracnose), and under normal conditions. The results showed distinct differences in the image entropy values between virus-infected, fungal-infected, and non-infected leaves, especially from NUV images. This indicates that the system effectively utilizes NIR and NUV imaging to detect diseases, with texture and image entropy analysis serving as reliable metrics. Notably, the system is more effective at early detection of fungal infections (such as anthracnose) than virus infections (such as Gemini virus), with NUV imaging proving more effective than NIR for this purpose.