A New Hybrid Intelligent GAACO Algorithm for Automatic Image Segmentation and Plant Leaf or Fruit Diseases Identification Using TSVM Classifier

Abstract

The aim of image segmentation process is to divide a digital image into sets of pixels. Image segmentation can play an important role not only in image segmentation but also in plant leaf or fruit disease detection. In this paper, we propose a new hybrid intelligent algorithm (GAACO) including Genetic Algorithm (GA), Ant Colony Optimization Algorithm (ACO) and Tabu list for different types of images segmentation as well as plant leaf or fruit image segmentation and transductive support vector machine (TSVM) is used to detect diseases of plant leaf or fruit. In this process, Genetic Algorithm is used to search for most optimal cluster centers in the problem space and then the Ant Colony Optimization is employed to achieve the best solution. Tabu list is used to save the image pixels into the memory. After image segmentation, the transductive support vector machine is used in testing phase and the obtained testing samples are compared with training samples. However, plant leaf disease or fruit disease detection is done by TSVM in accordance with leaf or fruit feature extraction. The result of the proposed algorithm shows that the hybrid GAACO algorithm gives high performance with a very low computational complexity, helps to enhance segmentation accuracy and supports TSVM to find the accurate diseases.