Piper betle and piper sarmentosum identification using support vector machine

Plant identification is an important task for biologists, chemists, scientists and environmentalists to identify every plant species in the world for environmental protection. However, the process of plants identification is usually very time consuming and low efficiency to identify between plants that are almost identical to each other especially in terms of shape. The objectives of this study are to extract the features properties and to identify the Piper betle and Piper sarmentosum using Support Vector Machine (SVM). These two types of plants are identical to each other because they come from the same family that is Piperaceae. The image of Piper betle and Piper sarmentosum leaves are used as the data to extract their features using image processing techniques. The features extraction properties from the leaf such as area, perimeter, convex hull, major axis, minor axis and the ratio of axes are calculated and extracted using MATLAB. The classification method of SVMused the extracted features to find the best possible hyperplane that can identify and classify between these two types of plants. Besides, the optimization value of the box constraint and kernel scale also been measured automatically by using the function code in image processing toolbox. The result demonstrated that the successful rate of plant identification of Piper betle and Piper sarmentosum is up to 80% using SVM method. As conclusion, the used of SVM is beneficial in classifying between Piper betle and Piper sarmentosum although they are almost identical in shape.