Font Size: a A A

Design And Implementation Of Vehicle Classification System Based On Embedded System

Posted on:2011-02-22Degree:MasterType:Thesis
Country:ChinaCandidate:H Z LiangFull Text:PDF
GTID:2178330332461018Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
With the rapid development of modern society economy and greatly improved quality of life, people begin the pursuit of higher level. So the traffic problems which are closely related people's life get more and more attention in this context. Therefore, intelligent transportation comes into being. It is a new transport system, which combines many advantages such as real-time, accuracy and efficient. Currently it is used more extensively in European and American and our country has started on it.In the field of intelligent transportation, vehicle type recognition is an attractive focus of attention for many researchers, because it is essential in achieving the automatic highway toll collection system and the practical application of intelligent parking system, it can be said to be a player key role. In practical applications, similar systems have higher requirements for real-time and identification accuracy, but unfortunately, due to various reasons, this have not got the promotion and popularity in domestic. After referring to a variety of vehicle types identification systems that exist today, adhering to use modern advanced science and technology and attitude of survival of the fittest,this paper designs and realizes the vehiele type classification based on embedded system. The system makes full use of the embedded system cost and resource advantages and finished the basic functions with emphasis on improving the kinds of vehicle identification accuracy. This system not only realizes a typical BP neural network, clustering and support vector machine in terms of vehicle type recognition on the applications, but also adopts the voting system inspired by multi-class support vector machine. It gets the final output by the results of the three output comprehensive judgments. The experiments tell that this kind of voting system improves vehicle classification accuracy and it complete with initial expectations.This paper introduces in detail the design process and technical realization, describes the hardware structure framework, introduces various vehicle type identification algorithms and implementations, carries on the final test and proves the system meets the initial design goals.
Keywords/Search Tags:Intelligent transportation, BP neural network, clustering, support vector machine, Vehicle Identification
PDF Full Text Request
Related items