Revolutionizing Leaf Classification for Agricultural Sustainability and Plant Species Analysis

Agricultural production is critical for economic stability and food security, consequently early disease identification is essential for preserving plant health. Our study solves this need by incorporating Random Forest into a complete framework for leaf categorization. We recommend applying Particle Swarm Optimization (PSO) to construct an AI-driven, four-step procedure for this type of segregation. Crop images may be processed by preprocessing, segmenting, extracting features, and classifying them with Random Forest. During pre-processing, a Median Filter was applied, proceeded by thresholding for segmentation, eventually Kaze and HOG (Histogram of Oriented Gradient) were used to extract relevant features. A Random Forest classifier that has been optimized with PSO was used for classification. The random forest-based technique outperformed alternative methods such as naive-bayes and adaboost with an accuracy of 92.63