Leaf Feature-Based Plant Spotting System with Random Forest Classification

Abstract

Identification and classification of unknown plant species are performed manually by expert personnel who are very few. The important aspect is to develop a system which classifies the plants. This paper presents a new recognition approach based on Leaf Features Fusion and Random Forests (RF) Classification algorithms for classifying the different types of plants. The proposed approach consists of three phases that are preprocessing, feature extraction, and classification phases. Since most types of plants have unique leaves. Leaves are different from each other by characteristics such as the shape, color, texture and the margin. This paper presents a classification approach by Random Forest Classifier algorithm for classifying the different types of plants. This Proposed Approach consists of four phases that are image pre-processing, feature extraction, features fusion and classification phases. Most types of plants have unique leaves. There are many features of leaf such as Color features, Vein features, GLCM features, Shape features and Gabor features. Also calculate Zernike moments such as amplitudes and phase. These all features are fused by concatenating of two vectors. So, the classification approach presented in this research depends on plant’s leaves. Experimental results show accuracy and other parameters measured in this approach with fusion of all these features or their different combinations. This is an intelligent system which can identify tree species from photographs of their leaves, and it provides accurate results in less time. “