【摘要】：近年來(lái),計(jì)算機(jī)視覺(jué)技術(shù)逐漸成熟,其在智能監(jiān)控領(lǐng)域的應(yīng)用愈加廣泛。大量原本需要人工完成的工作都可以由視覺(jué)算法來(lái)替代,極大地節(jié)約了人力成本。而在智能監(jiān)控領(lǐng)域中,行人流量統(tǒng)計(jì)這一技術(shù)在商場(chǎng)、校園等場(chǎng)合都有著廣泛的需求和應(yīng)用,因此設(shè)計(jì)出一套智能的行人流量統(tǒng)計(jì)算法是十分有必要的。另一方面,如果能夠?qū)⒔谂d起的深度學(xué)習(xí)技術(shù)應(yīng)用于其中,則將極大地提升算法的性能。本文研究并設(shè)計(jì)了一種基于深度學(xué)習(xí)的行人流量統(tǒng)計(jì)算法。本文通過(guò)綜合應(yīng)用基于深度學(xué)習(xí)的目標(biāo)檢測(cè)算法、單目標(biāo)跟蹤算法、數(shù)據(jù)關(guān)聯(lián)算法等方法,設(shè)計(jì)了一套框架為“檢測(cè)-跟蹤-關(guān)聯(lián)”的算法,用來(lái)完成對(duì)監(jiān)控視頻中行人流量的統(tǒng)計(jì)。本文主要進(jìn)行了以下研究工作:首先,本文探究了行人流量統(tǒng)計(jì)的應(yīng)用背景并闡述了研究的意義,然后分析了行人流量統(tǒng)計(jì)技術(shù)和基于深度學(xué)習(xí)的目標(biāo)檢測(cè)算法的發(fā)展現(xiàn)狀,接下來(lái)闡述了本文的主要研究?jī)?nèi)容和研究方案,在研究方案中給出設(shè)計(jì)好的總體算法和框架。在制定了較為完善的研究方案的基礎(chǔ)上,本文首先研究了卷積神經(jīng)網(wǎng)絡(luò)的組成結(jié)構(gòu)和優(yōu)化方法。然后回顧了傳統(tǒng)目標(biāo)檢測(cè)算法和近年來(lái)產(chǎn)生的基于卷積神經(jīng)網(wǎng)絡(luò)的目標(biāo)檢測(cè)算法,最后確定使用SSD算法作為行人流量統(tǒng)計(jì)中的目標(biāo)檢測(cè)算法。接下來(lái),本文研究了SSD算法的框架與原理,包括網(wǎng)絡(luò)結(jié)構(gòu)、缺省框的選擇和訓(xùn)練目標(biāo)函數(shù)等。之后,重點(diǎn)研究了SSD中的基網(wǎng)絡(luò),參照SSD原始基網(wǎng)絡(luò)VGG和流行的CNN網(wǎng)絡(luò)結(jié)構(gòu)ZF-Net和SqueezeNet,重新設(shè)計(jì)了兩種基網(wǎng)絡(luò)并與VGG進(jìn)行比較,結(jié)合實(shí)際需求最終確定了還是使用VGG作為基網(wǎng)絡(luò)。在完成了對(duì)基于卷積神經(jīng)網(wǎng)絡(luò)的檢測(cè)算法的研究后,本文還研究了需要使用的跟蹤算法、數(shù)據(jù)關(guān)聯(lián)算法和軌跡分析算法。確定了使用KCF算法作為跟蹤算法,并直接使用OpenCV中的跟蹤庫(kù)。關(guān)聯(lián)算法選取簡(jiǎn)單快速的基于距離的關(guān)聯(lián)算法。最后設(shè)計(jì)了軌跡分析算法來(lái)實(shí)現(xiàn)雙方向的計(jì)數(shù)。完成上述工作后,算法便已經(jīng)完整。最后,闡述了實(shí)際的操作,包括攝像頭的架設(shè)和樣本視頻的采集、檢測(cè)圖像數(shù)據(jù)集的制作、SSD檢測(cè)器的訓(xùn)練、跟蹤算法的實(shí)現(xiàn)、關(guān)聯(lián)與軌跡分析算法的設(shè)計(jì)要點(diǎn)。然后使用設(shè)計(jì)的算法對(duì)所有樣本視頻進(jìn)行分析,采用一些性能指標(biāo)對(duì)其表現(xiàn)進(jìn)行評(píng)價(jià),其中平均識(shí)別率達(dá)到了96.24%,平均誤檢率為2.19%,平均漏檢率為3.76%,全部視頻平均幀率為24.09。結(jié)果表明所設(shè)計(jì)的算法可以能夠滿足項(xiàng)目的需求。
[Abstract]:In recent years, computer vision technology is gradually mature, its application in the field of intelligent monitoring is becoming more and more extensive. A large amount of manual work can be replaced by visual algorithm, which greatly saves manpower cost. In the field of intelligent monitoring, pedestrian flow statistics technology has a wide range of needs and applications in shopping malls, campus and other occasions, so it is necessary to design a set of intelligent pedestrian flow statistics algorithm. On the other hand, if the recently developed depth learning technology can be applied to it, the performance of the algorithm will be greatly improved. In this paper, a pedestrian flow statistic algorithm based on depth learning is studied and designed. In this paper, a set of algorithms called "detection, tracking and association" is designed by synthesizing the methods of target detection based on depth learning, single target tracking and data association. Used to complete the monitoring video traffic statistics. The main work of this paper is as follows: first, this paper explores the application background of pedestrian flow statistics and expounds the significance of the research, and then analyzes the development status of pedestrian flow statistics technology and target detection algorithm based on depth learning. Then, the main research content and research scheme of this paper are described, and the overall algorithm and framework are given in the research scheme. On the basis of a more perfect research scheme, this paper first studies the composition structure and optimization method of convolution neural network. Then the traditional target detection algorithm and the target detection algorithm based on convolution neural network are reviewed. Finally, the SSD algorithm is used as the target detection algorithm in pedestrian flow statistics. Then, this paper studies the framework and principle of SSD algorithm, including network structure, selection of default frame and training objective function. After that, the base network in SSD is studied emphatically. Referring to the original SSD network VGG and the popular CNN network structure ZF-Net and SqueezeNet, two base networks are redesigned and compared with VGG. Finally, VGG is used as the base network according to the actual requirements. After completing the research on the detection algorithm based on convolution neural network, this paper also studies the tracking algorithm, data association algorithm and trajectory analysis algorithm that need to be used. The KCF algorithm is used as the tracking algorithm, and the trace library in OpenCV is used directly. The association algorithm selects the simple and fast distance based association algorithm. Finally, a trajectory analysis algorithm is designed to realize double direction counting. After the above work has been completed, the algorithm is complete. Finally, the practical operation, including the installation of camera and the collection of sample video, the training of SSD detector, the realization of tracking algorithm and the design of correlation and trajectory analysis algorithm are discussed. Then the designed algorithm is used to analyze all the sample video, and some performance indexes are used to evaluate the performance of the video. The average recognition rate is 96.24, the average false detection rate is 2.19, the average missed detection rate is 3.76, and the average frame rate of the whole video is 24.09. The results show that the algorithm can meet the requirements of the project.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 韓娜;陳東偉;鐘卓成;;智能視頻客流人數(shù)統(tǒng)計(jì)系統(tǒng)的算法比較研究[J];信息技術(shù);2016年06期

2 肖江;丁亮;束鑫;張文章;;一種基于計(jì)算機(jī)視覺(jué)的行人流量統(tǒng)計(jì)方法[J];信息技術(shù);2015年08期

3 徐超;高夢(mèng)珠;查宇鋒;曹利民;;基于HOG和SVM的公交乘客人流量統(tǒng)計(jì)算法[J];儀器儀表學(xué)報(bào);2015年02期

4 潘泓;李曉兵;金立左;夏良正;;一種基于二值粒子群優(yōu)化和支持向量機(jī)的目標(biāo)檢測(cè)算法[J];電子與信息學(xué)報(bào);2011年01期

5 李娟;楊錄;張艷花;;基于紅外傳感器的樓宇人數(shù)統(tǒng)計(jì)系統(tǒng)的設(shè)計(jì)[J];山西電子技術(shù);2010年06期

6 葉林;陳岳林;林景亮;;基于HOG的行人快速檢測(cè)[J];計(jì)算機(jī)工程;2010年22期

7 田牛;應(yīng)捷;;基于Canny邊緣幀差法的公交客流量統(tǒng)計(jì)[J];微計(jì)算機(jī)信息;2010年32期

8 周晨卉;王生進(jìn);丁曉青;;基于局部特征級(jí)聯(lián)分類器和模板匹配的行人檢測(cè)[J];中國(guó)圖象圖形學(xué)報(bào);2010年05期

9 余莉;韓方劍;;基于分水嶺變換和遺傳算法的自動(dòng)目標(biāo)檢測(cè)[J];中國(guó)圖象圖形學(xué)報(bào);2008年09期

10 李同治;丁曉青;王生進(jìn);;利用級(jí)聯(lián)SVM的人體檢測(cè)方法[J];中國(guó)圖象圖形學(xué)報(bào);2008年03期

相關(guān)碩士學(xué)位論文 前10條

1 李娟;基于KCF的視頻中運(yùn)動(dòng)物體的跟蹤系統(tǒng)[D];湖南師范大學(xué);2016年

2 蔡澤彬;基于視頻分析的行人檢測(cè)及統(tǒng)計(jì)方法研究[D];華南理工大學(xué);2015年

3 g罘，

本文編號(hào)：2240506