A systematic review of hyperspectral imaging in precision agriculture: Analysis of its current state and future prospects

Abstract

Hyperspectral sensor adaptability in precision agriculture to digital images is still at its nascent stage. Hyperspectral imaging (HSI) is data rich in solving agricultural problems like disease detection, weed detection, stress detection, crop monitoring, nutrient application, soil mineralogy, yield estimation, and sorting applications. With modern precision agriculture, the challenge now is to bring these applications to the field for real-time solutions, where machines are enabled to conduct analyses without expert supervision and communicate the results to users for better management of farmlands; a necessary step to gain complete autonomy in agricultural farmlands. Significant advancements in HSI technology for precision agriculture are required to fully realize its potential. As a wide-ranging collection of the status of HSI and analysis in precision agriculture is lacking, this review endeavors to provide a comprehensive overview of the recent advancements and trends of HSI in precision agriculture for real-time applications. In this study, a systematic review of 163 scientific articles published over the past twenty years (2003–2023) was conducted. Of these, 97 were selected for further analysis based on their relevance to the topic at hand. Topics include conventional data preprocessing techniques, hyperspectral data acquisition, data compression methods, and segmentation methods. The hardware implementation of field-programmable gate arrays (FPGAs) and graphics processing units (GPUs) for high-speed data processing and application of machine learning and deep learning technologies were explored. This review highlights the potential of HSI as a powerful tool for precision agriculture, particularly in real-time applications, discusses limitations, and provides insights into future research directions.