A Shape and Texture Approach for Plant Identification

Plants are the foremost important resources on earth as they supply raw materials for basic needs, medicines and many other products that help to sustain our lives. Plants recognition has become a hindrance especially for biologists, chemists and environmentalists. People lack basic knowledge about the plant species that have health benefits and provide food to the mankind. An automated plant identification system can be castoff for rapid categorization of plant species without the need of expertise, thus atomizing their task. The method proposed is used to identify the leaf images that consist of certain phases. Firstly pre-processing technique is applied to enhance the leaf images. In the next stage, leaf features are extracted on basis of shape and texture features. Further fusion of features was incorporated and fed to two classifiers like SVM(Support Vector Machine) and KNN(k-nearest neighbors ) for leaf identification. The shape and texture features achieved an accuracy of 81.1% and 80% with SVM and KNN when tested with dataset that contains 100 images of different plant leaves.