A Vehicle Classification System Based On Inductive Loop Detector

Pattern recognition is one of the most important regions of Artificial Intelligence (A.I). It is in order to research complicated process of dealing with information. With the development of computer technology, pattern recognitionisusedtoA.I, computer engineering, robot, nerve-biology e tc widely, which basic theory and method are more richer. The paper is based on the basic theory of pattern recognition, and doing deeply research and discussion about the automatic recognition of automobile type.At present, the automatic recognition of automobile type is one of Intelligent Transportation System (ITS) hot topics, which is the core of the charge system that does not need stop vehicle, and has practice meaning in future. The thesis content came from VATA Ltd. request, which is a vehicle classification system based on inductive loop detectors, made up with data collection sample wave character extracting and automobile type pattern matching.This paper introduces principle of data collection and hardware frame, and examples some common automobile type of data ware at the same time. After analyzing data and comparing automobile type of data ware, the max-data as criterion of vehicle classification, whole sample's data as character parameter of pattern matching.Because of different velocity through inductive loop detector, sample ware has flexibility. When doing automobile type pattern matching, in order to keep similar character consistent, the paper put forward the project that symmetry DTW and non- symmetry DTW use together. By using MATLAB software, I validate feasibility of this arithmetic, and obtain anticipate effect.