Abstract

Complete environment system depends on various plants. Plants help to maintain ecological system by controlling pollution and providing oxygen etc. but due to inappropriate information about plant may lead to decreasing the number of plant species resulting in imbalance in ecological system. Hence an automated system is required for botanists to categorize plants for maintains the database for further researches. In this field, computer vision techniques have proven significant techniques. For plant classification, an automated system is presented using improved pre-processing (by improving image enhancement or filtering scheme). Furthermore, wavelet feature and local features are also combine for feature vector formulation. Finally, a comparative classification study is presented where KNN, Neural Network and SVM classifiers are implemented. experimental study shows that SVM provides better classification performance and further performance can be improved by using combined feature extraction model.