Index for Efficient Pest Detection Using Multispectral Imaging

Abstract Pesticides have many benefits for agriculture but do pose a problem when unmanaged. Pests can be difficult to detect and if pest activities such as growth and travel patterns can be easily recorded, then there can be an early detection of any potential problems such as overgrowth that can lead to environmental hazards. Another important benefit of detection methods for pests is the detection of invasive species that affect agriculture by decimating crops, causing economic losses and disruptions in food supply chains, especially during cross border activities. In this work, multispectral imaging is used to develop an index for pest detection using a monochromatic light source. A selected pest such as beetles is used in different environments. A new index would be able to capture spectral patterns or signatures associated with pest presence/damage with a high accuracy. The intended research can be expanded for many different pest species. The experimental setup is made with different illumination sources including a monochromatic red light and diffused light. Using a multispectral camera, a reflectance analysis is conducted from the images acquired, and an index indicating the presence of pests in rice is found and proposed herein. Ultimately, this study can aid in reducing the risk of invasive pests entering the US borders undetected, which can improve the food supply chains and strengthen the nation’s economy.