基于信息融合的车载视频变流量诱传决策算法研究

发布时间：2018-03-24 19:34

  本文选题：车载视频　切入点：变流量传输　出处：《浙江工业大学》2014年硕士论文

【摘要】：随着社会经济和科技的不断发展,公路运输在给人类生活带来便利的同时,影响交通运输安全的恶性事件时有发生,公路交通治安问题俨然成为公路运输业和公安系统的难题。视频作为信息载体,已成为公路交通监控中的一种主要方式。然而当前基于移动公网的信息传输流量费用仍然较高,本文研究基于信息融合的车载视频信息变流量传输算法,可有效降低昂贵的流量费用,主要工作和成果如下：(1)分析信息融合和BP (Back-Propagation,反向传播)神经网络的优缺点,针对BP网络的输入信息复杂多变时,固定的网络结构不能适应各种变化的环境的问题,提出了结合DS (Dempster-Shafe r证据理论)和BP神经网络的决策算法,即DS-]BP决策算法。(2)通过研究车辆运行异常的原因,对车辆的视频、音频、车速、加速度、转向强度、发动机转速和冷却液温度等信息进行异常特征提取,为之后的车载视频变流量传输决策奠定数据基础。(3)利用PSO (Particle Swarm Optimization,粒子群优化)算法来确定BP神经网络的初值,改善了BP神经网络容易陷入局部极小值和收敛速度慢的问题。
[Abstract]:With the continuous development of social economy and science and technology, highway transportation brings convenience to human life, and at the same time, the malignant events that affect the safety of transportation occur from time to time. The problem of highway traffic security has become a difficult problem in highway transportation and public security system. As a carrier of information, video has become a main way of highway traffic monitoring. However, the cost of information transmission based on mobile public network is still relatively high. This paper studies the variable traffic transmission algorithm based on information fusion, which can effectively reduce the high traffic cost. The main work and results are as follows: 1) analyzing the advantages and disadvantages of information fusion and BP Back-Propagation) neural network. In view of the problem that the fixed network structure can not adapt to various changing environments when the input information of BP network is complex and changeable, a decision algorithm based on DS Dempster-Shafe r evidence theory and BP neural network is proposed. That is, DS-] BP decision algorithm. 2) by studying the cause of the abnormal operation of the vehicle, the abnormal features of the video, audio, speed, acceleration, steering strength, engine speed and coolant temperature of the vehicle are extracted. This paper establishes the data base for the later decision of vehicle video variable traffic transmission. It uses PSO / Particle Swarm optimization (PSO) algorithm to determine the initial value of BP neural network, which improves the problem that BP neural network is easy to fall into local minimum value and slow convergence speed.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TP202
，

本文编号：1659725