Identifikasi Plat Nomor Menggunakan Fitur Zoning dengan Klasifikasi Support Vector Machine
Abstract
Vehicle identification detection is one of the significant problems with the increasing number of vehicles. Therefore, a computer-based method is needed that can identify the vehicle based on license plate numbers quickly and accurately. Previous research have applied the image centroid and zone (ICZ) feature extraction method to identify vehicle license plates. In this research, ICZ and support vector machine (SVM) will be used for license plate identification. SVM which is used is the multi class SVM one against all using linear kernel, the polynomial, and RBF. The testing is performanced twice, on each character and on the overall plate (with or without fault tolerance). From the three kernels, the kernel which produces the best accuracy is the polynomial kernel with a value of C equals to 0.125 and d equals to 2 with on accuracy of 95.44%, while the accuracy produced at plate testing without fault tolerance is 81.54% and testing with fault tolerance equal to 1 is 90.77%
Collections
- UT - Computer Science [2322]