Computer Vision for Plant Disease Recognition: A Comprehensive Review

Abstract

Agriculture has undergone a remarkable transformation, transitioning from traditional methods that were used for centuries to technology-driven practices. The advent of image processing and computational intelligence has revolutionized crop production and plant health monitoring. From drones capturing detailed crop growth data to sensors meticulously measuring soil moisture levels, the possibilities are boundless. This review delves into the cutting-edge research advancements in the application of image processing and computational intelligence techniques for botanical fields, with a particular focus on plant health monitoring. First, it provides a comprehensive overview of the diverse imaging sensors employed in agriculture, including visible, near-infrared, thermal, and hyperspectral imaging. Subsequently, it carefully analyzes the advantages and limitations of each sensor type, along with illustrative examples of their utilization in plant health monitoring. The review further explores the application of machine learning and deep learning for automated plant disease identification, highlighting the critical need for standardized datasets, benchmarking protocols, and domain-specific knowledge for effective implementation. In conclusion, the review emphasizes the future challenges and trends in this rapidly evolving field. It serves as a valuable resource, providing insights into the latest trends in computer vision-based plant disease monitoring and identifying gaps that demand further attention from the scientific community.