Plant leaf identification of kaffir lime leaves using Support Vector Machine (SVM)

Plant identification is very important in day-to-day life especially for biologists, chemists, and scientists to identify every plant species and it is interesting research due to various types of plant species. Numerous studies have been conducted for plant leaf identification especially when the plants have various types of leaf and the plants leaves are almost identical to each other. For this study, two types of different shapes of kaffir lime leaf, one is a figure eight-shape of double leaves and the other one is shaped wide at the top and slightly small at the base has been chosen as the subject of the study. This study focuses on the feature extraction of kaffir lime leaves such as area, parameter, convex hull, major axis, minor axis and the ratio of major and minor axis and determination of the best accuracy between three (3) different kernel functions which are the radial basis function (rbf), linear and polynomial kernel for plant identification of two types of kaffir lime leaves using Support Vector Machine (SVM). A machine learning algorithm, SVM, has been used to create species identification models. The 20 leaves images of figure eight-shape of double leave and 20 leaves images of shape wide at the top and slightly small at the base of the kaffir lime plant were used as the data to determine the features extraction using MATLAB R2021a software. Besides that, the optimization value of the box constraint for three (3) different kernel functions also have been measured automatically by using the function code in the image processing toolbox. The result demonstrated that the most accurate rate of plant identification of two types of kaffir lime leaves is up to 91.67% by using linear kernel function in SVM method.