Affiliations: [a] Faculty of Science, Al-Azhar University, Cairo, Egypt | [b] Machine Intelligence Research Labs, MIR Labs, Auburn, WA, USA
Correspondence:
[*]
Corresponding author: Heba F. Eid, Faculty of Science, Al-Azhar University, Cairo, Egypt. E-mail: [email protected].
Abstract: The classification of plants species is a crucial process in some agricultural-based industries. However, different plant species share a very close relationship to human beings. This paper proposes a plant identification model based on leaf biometrics (shape, texture and color) hybrid with two most recent swarm optimization algorithms. For which, particle Swarm Optimization (PSO) is adopted as a pre-processing phase for leaf image segmentation. While, Grey Wolf Optimizer (GWO) is obtained to reduce the dimension of the leaf texture descriptors. Finally, the dual coordinate descent L2-SVM classifier is used to classify the different plant species. The proposed model aims to achieve high identification accuracy using less leaf’s descriptors. Several experiments on Flavia dataset and swedish dataset are conducted. The experimental analysis showed that, the proposed model yields to improve the identification rate up to 98.9% and 93.3% for both Flavia and Swedish dataset respectively, which are the improved values over the literature.
Keywords: Plant identification, bio-inspired optimization, Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), support vector machine (SVM)
