Automated Plant Species Identification Using Leaf Shape-Based Classification Techniques: A Case Study on Iranian Maples

Foliar characteristics, especially the overall leaf shape, are useful features for the taxonomic identification of plants. Computer-aided plant species identification systems make it possible to investigate a large number of leaves in a short period of time. In this study, a fully automatic system was developed to accurately classify eight species of maples (Acer L.) in Iran using the leaf shape characteristics of harvested leaves. Maples show a broad range of leaf morphology, and the provided dataset is a diverse collection of simple leaves which can be a good representative of woody plant leaves with any overall shape and margin pattern. The applied method consisted of preprocessing, feature extraction and classification steps. A new method for petiole removal and a new feature as teeth indicator are presented in preprocessing and feature extraction parts, respectively. Using support vector machine (SVM) classifier, experimental results showed a high accuracy of 98.60% ± 1.20 for the identification of maple species. Focusing on extracting robust features, the proposed method offers considerable potential to speed up expert identification of maples with high accuracy whose output can be applied to computational botany, conservation planning and natural resource management.