Particle Swarm Optimization (PSO) with fuzzy c means (PSO-FCM)–based segmentation and machine learning classifier for leaf diseases prediction

摘要

This paper proposes an automatic classification technique that uses leaf images some medicinal plants. It is primarily the core reason that drives the research presented here, including the introduction of new innovative segmentation and classification techniques that are deployed to facilitate automatic detection. The major aim of the work is to introduce a new leaf disease prediction technique. The study conducted here a unique but effective image segmentation, feature extraction, as well as plant leaf disease classification. The proposed approach initially preprocesses leaf images of plants thereafter which the diseased sections of the plant are segmented by deploying Particle Swarm Optimization (PSO)-based fuzzy c means segmentation (PSO-FCM), Gaussian Mixture Model (GMM)-based background subtraction. Vein and shape features, edge-based feature extraction, and texture characteristics or texture features (TF) are computed. This methodology classifies the leaves of medicinal plants by deploying the Multiple Kernel Parallel Support Vector Machine (MK-PSVM) classifier. The classifier is implemented via the use of MATLAB classifier. The results are measured using the accuracy, sensitivity, specificity, precision, and F-measure metrics. Experimental results depict that the classifiers that have been proposed here achieve a higher classification accuracy enabling leaf detection.