HYBRID CLASSIFICATION METHOD FOR WILD OLIVE BASED ON GLCM AND NEURAL NETWORKS IN SYRIAN REGIONS

The recognition and identification of plant species are very time-consuming as it has been mainly carried out by botanists. The texture is one of the most popular features used for image classification. In this paper we propose a hybrid method for clustering wild olive trees which has been designed and developed to recognize typical texture features for olive leaves digital images. The textures’ features extracted from gray level co-occurrence matrices (GLCM) are the typical values for features analysis in classification. The proposed method is tested on a data base of 210 images leaves with 14 images for each variety (class). The experiments were accomplished by using 15 types of wild olive trees. An artificial neural network has been used to classify pairs of two types. We obtained an accuracy matrix for the classification rate over all types. The obtained accuracy for each pair is considered as a distance between the pairs. Based on the accuracy matrix and the Unweighted Pair Group Method Centroid (UPGMC), we constructed the clustering tree for the 15 wild olive types. The preliminary results obtained indicate the technical feasibility of the proposed method, which will be applied for more varieties from Syria.