Car model recognition from frontal image using BRISK

摘要

Car Recognition is a part of Intelligent Transportation System. This research proposes the manufacture of ITS-based system to identify car model from Its frontal image using Binary Robust Invariant Scalable method. The BRISK method is used to detect image keypoint, and it uses Hamming Distance for keypoint matching. As for matching error, this research depends on RANSAC. BRISK method excellence lies in the speed of scale space searching of images. Car model recognition using BRISK has an accuracy rate of 96.25%.