局所特徴認識アルゴリズムによる車両の認識

Abstract

In this paper, we propose a robust method for recognizing vehicles. Our system is based on local-feature configuration method, which is a generalization of eigen-window method. This method has following three advances: (1) It can detect even if part of vehicle is occluded. (2) It can detect even if vehicles are translated due to the running out of the lanes. (3) It does not require us to segment vehicle areas from input images. We have developed our system with infrared images, but in this paper, we apply our system on images of super wide-angle optical images. In outdoor experiments, our system detects over 95 % of vehicles.