本文選題：多車(chē)型信息采集　切入點(diǎn)：視頻檢測(cè)　出處：《長(zhǎng)安大學(xué)》2014年碩士論文　論文類(lèi)型：學(xué)位論文

【摘要】：準(zhǔn)確的道路交通流信息采集和符合中國(guó)國(guó)情的交通流基礎(chǔ)理論研究已經(jīng)成為我國(guó)發(fā)展智能交通系統(tǒng)（IntelligentTransportSystem，ITS）緩解交通壓力減少交通事故以及提高道路通行能力的的關(guān)鍵所在隨著計(jì)算機(jī)技術(shù)的飛速發(fā)展，視頻分析技術(shù)越來(lái)越受到研究人員的追捧本文在分析各種交通流信息采集方法的基礎(chǔ)上，設(shè)計(jì)了一套多場(chǎng)景條件下分車(chē)型交通信息視頻采集系統(tǒng)，并在VS2010環(huán)境下配置Opencv（OpenSourceComputerVisionLibrary，開(kāi)源計(jì)算機(jī)視覺(jué)庫(kù)）最新版2.4.4版本開(kāi)發(fā)了相應(yīng)軟件 文章介紹了常用的幾種交通信息采集方法，闡述了視頻信息采集的優(yōu)勢(shì)和現(xiàn)狀，對(duì)視頻檢測(cè)技術(shù)方面的發(fā)展現(xiàn)狀和發(fā)展趨勢(shì)進(jìn)行了具體分析；詳細(xì)介紹了所開(kāi)發(fā)系統(tǒng)的操作流程，并對(duì)其進(jìn)行了效果展示；重點(diǎn)介紹了系統(tǒng)開(kāi)發(fā)過(guò)程中用到的關(guān)鍵技術(shù)：利用彩色視頻提取初始背景提高了幀差法的準(zhǔn)確度，利用基于時(shí)間的圖像疊加技術(shù)更新背景來(lái)達(dá)到自適應(yīng)背景更新效果，采用二值圖像形態(tài)學(xué)處理來(lái)補(bǔ)充背景差分造成的運(yùn)動(dòng)車(chē)輛信息丟失，，選用直接線性變換方法實(shí)現(xiàn)了像素坐標(biāo)與世界坐標(biāo)的轉(zhuǎn)換，采取 少數(shù)服從多數(shù)的投票法提高車(chē)型以及車(chē)輛瞬時(shí)速度的測(cè)定準(zhǔn)確度，采用設(shè)置參數(shù)外置的方式實(shí)現(xiàn)了多場(chǎng)景下分車(chē)型的交通信息參數(shù)檢測(cè)；同時(shí)系統(tǒng)給出簡(jiǎn)便的系統(tǒng)參數(shù)設(shè)置方法和檢測(cè)結(jié)果的保存方式，除此之外對(duì)實(shí)際視頻進(jìn)行速度車(chē)型流量檢測(cè)并對(duì)結(jié)果進(jìn)行了分析，指出交通流的特性和系統(tǒng)中可能存在的問(wèn)題 測(cè)試結(jié)果顯示系統(tǒng)無(wú)論在車(chē)速車(chē)型還是流量檢測(cè)方面的檢測(cè)準(zhǔn)確率均滿足理論研究分析需要在此基礎(chǔ)上對(duì)西安市西三環(huán)南二環(huán)和東二環(huán)采集到的交通視頻數(shù)據(jù)結(jié)果進(jìn)行理論分析，按照車(chē)型類(lèi)別討論其參數(shù)變化規(guī)律并對(duì)檢測(cè)結(jié)果中蘊(yùn)含的交通流原理進(jìn)行了解析
[Abstract]:Accurate information collection of road traffic flow and basic theory research of traffic flow in accordance with China's national conditions have become the key to the development of Intelligent Transport system (ITSs) in China to alleviate traffic pressure and reduce traffic accidents and improve road traffic capacity. With the rapid development of computer technology, Video analysis technology is becoming more and more popular by researchers. Based on the analysis of various traffic flow information collection methods, this paper designs a multi-scene model traffic information video acquisition system. The corresponding software is developed in the latest version 2.4.4 of Open Source computer Vision Library (Open Source computer Vision Library) in VS2010 environment. This paper introduces several common methods of traffic information collection, expounds the advantages and present situation of video information collection, and analyzes the development status and development trend of video detection technology. The operating flow of the developed system is introduced in detail, and its effect is demonstrated. The key technologies used in the development of the system are emphatically introduced: the accuracy of the frame difference method is improved by using color video to extract the initial background. The time-based image superposition technique is used to update the background to achieve the adaptive background updating effect, and the binary image morphology processing is used to supplement the motion vehicle information loss caused by background difference. The direct linear transformation method is used to realize the transformation between pixel coordinates and world coordinates. A few clothes from the majority of voting method to improve the vehicle and vehicle instantaneous speed measurement accuracy, using the setting of parameters outside the way to achieve the multi-scene sub-vehicle traffic information parameters detection; At the same time, the system gives a simple method of setting the system parameters and saving the test results. In addition, the speed and vehicle flow detection of the actual video is carried out and the results are analyzed. The characteristics of the traffic flow and the possible problems in the system are pointed out. The test results show that the detection accuracy of the system in both vehicle speed and flow detection meets the needs of theoretical research and analysis. On this basis, traffic video data collected from South second Ring and East second Ring of Xi'an City are collected. The results were analyzed theoretically. According to the type of vehicle, the variation law of its parameters is discussed, and the principle of traffic flow contained in the detection results is analyzed.
【學(xué)位授予單位】：長(zhǎng)安大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類(lèi)號(hào)】：U491.116

【參考文獻(xiàn)】

相關(guān)期刊論文 前6條

1 蘇禮坤;黃繼紅;;基于幀間差分的彩色視頻背景提取[J];成都信息工程學(xué)院學(xué)報(bào);2010年02期

2 翟永前;于惠芬;趙力;鄒采榮;;一種基于視頻技術(shù)的車(chē)輛跟蹤方法[J];電氣電子教學(xué)學(xué)報(bào);2011年06期

3 金偉其;現(xiàn)代城市交通流量的三維圖像檢測(cè)技術(shù)[J];紅外與激光工程;1997年06期

4 胡煒煒;李樹(shù)廣;吳舟舟;;基于邊緣輪廓特征的實(shí)時(shí)車(chē)型分類(lèi)[J];計(jì)算機(jī)工程;2007年13期

5 盧s

本文編號(hào)：1625192