本文選題：異常事件檢測(cè) + 排斥力�。� 參考：《鄭州大學(xué)》2017年碩士論文

【摘要】：近年來,隨著越來越多人群踩踏、恐怖襲擊以及其他公共安全事件的發(fā)生,作為群體行為監(jiān)控、分析、預(yù)警基礎(chǔ)的人群異常事件檢測(cè)已經(jīng)成為智能監(jiān)控領(lǐng)域中亟需解決的問題之一。然而,由于人群運(yùn)動(dòng)復(fù)雜多變,要列出人群中所有可能出現(xiàn)的異常事件幾乎是不可能的,因此人群異常事件的檢測(cè)并不是一個(gè)典型的分類問題。為此本文采用先從只包含正常事件的人群視頻中訓(xùn)練出一個(gè)特征集,再通過計(jì)算待檢測(cè)視頻中人群特征與特征集的偏離程度來判斷人群中是否出現(xiàn)異常這一技術(shù)路線。具體地,本文將人群異常事件的檢測(cè)分為兩步:一是人群視頻特征提取,即事件表示;二是模型的訓(xùn)練更新和人群異常事件檢測(cè)。對(duì)于人群視頻的特征提取,本文采用了兩種不同方法。一是基于排斥力的特征提取。為避免在擁擠場景中跟蹤每個(gè)具體個(gè)體帶來的問題,本文使用一種粒子的平流運(yùn)動(dòng)來模擬人群的運(yùn)動(dòng),并將通過排斥力模型獲取的能夠準(zhǔn)確反映出人群運(yùn)動(dòng)的力流矩陣視為提取的特征。二是基于卷積神經(jīng)網(wǎng)絡(luò)的特征提取。雖然第一種方法提取的特征在多種類型場景的異常檢測(cè)中都取得了良好的效果,但是它對(duì)于排斥力變化不明顯的場景敏感度不高。因此我們采用一種改進(jìn)的卷積神經(jīng)網(wǎng)絡(luò)模型,并結(jié)合滑動(dòng)窗口和PCA以提取人群視頻的時(shí)空特征。對(duì)于模型的訓(xùn)練和異常事件檢測(cè),本文采用了一個(gè)基于稀疏編碼的分組詞典框架以改善單一模式詞典帶來的計(jì)算量大且難以維護(hù)的問題。在異常事件檢測(cè)時(shí),本文依次使用每個(gè)詞典對(duì)待檢測(cè)單詞進(jìn)行稀疏重構(gòu)。一旦待檢測(cè)單詞能被某個(gè)詞典稀疏表示,那么就認(rèn)為它所表示的事件是正常的;若所有的詞典都不能稀疏表示它,那么就認(rèn)為是異常的。為解決隨著新視頻的不斷增加和視頻場景的動(dòng)態(tài)變化而導(dǎo)致的詞典組表示能力退化或概念漂移問題,本文基于稀疏重構(gòu)和一組單詞池提出了一個(gè)無監(jiān)督的詞典組局部和全局在線更新算法。最后,本文在公開數(shù)據(jù)集UMN、UCSD以及Web數(shù)據(jù)集上與其他方法進(jìn)行了對(duì)比分析。實(shí)驗(yàn)表明本文的兩種方法都提高了人群異常事件檢測(cè)的準(zhǔn)確度和效率,并且基于卷積神經(jīng)網(wǎng)絡(luò)特征提取的方法改善了基于排斥力的特征提取方法對(duì)于排斥力變化不明顯場景敏感度不高的問題。
[Abstract]:In recent years, as more and more people stampede, terrorist attacks and other public safety incidents, as group behavior monitoring, analysis, Detection of abnormal events based on early warning has become one of the most urgent problems in intelligent monitoring field. However, due to the complexity of crowd movement, it is almost impossible to list all possible abnormal events in the crowd, so the detection of abnormal events is not a typical classification problem. In this paper, we first train a feature set from a crowd video containing only normal events, and then calculate the deviation between the crowd feature and the feature set to determine whether there is an anomaly in the crowd. Specifically, the detection of crowd abnormal events is divided into two steps: one is crowd video feature extraction, that is, event representation; the other is model training update and crowd abnormal event detection. For crowd video feature extraction, this paper adopts two different methods. One is the feature extraction based on repulsive force. To avoid the problem of tracking each individual in a crowded scenario, a particle advection motion is used to simulate the movement of a population. The force flow matrix obtained by repulsive force model can accurately reflect the movement of people is regarded as the feature of extraction. Second, feature extraction based on convolution neural network. Although the feature extracted by the first method has achieved good results in anomaly detection of many kinds of scenes, it is not sensitive to scenes where the repulsive force is not obvious. So we adopt an improved convolution neural network model and combine sliding window and PCA to extract temporal and spatial features of crowd video. For model training and anomaly event detection, a group dictionary framework based on sparse coding is used to improve the computational complexity and difficult maintenance of single pattern dictionary. In anomaly event detection, each dictionary is used for sparse reconstruction of detected words. Once a word to be detected can be represented sparsely by a dictionary, it is assumed that the event it represents is normal; if all dictionaries cannot represent it sparsely, then it is considered abnormal. In order to solve the problem of loss of ability or concept drift of dictionaries caused by the increasing number of new videos and the dynamic changes of video scenes, Based on sparse reconstruction and a set of word pools, an unsupervised online updating algorithm for local and global dictionary groups is proposed in this paper. Finally, this paper compares and analyzes the UMN / UCSD and Web datasets with other methods. The experiments show that the two methods in this paper improve the accuracy and efficiency of the detection of crowd abnormal events. And the feature extraction method based on convolution neural network improves the problem that the feature extraction method based on repulsive force is not sensitive to the change of repulsion force.
【學(xué)位授予單位】：鄭州大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41;TN948.6

【相似文獻(xiàn)】

相關(guān)期刊論文 前6條

1 申喜芹;郭瓊瓊;張宸豪;彭宏玉;;高速公路異常事件智能響應(yīng)系統(tǒng)設(shè)計(jì)[J];中國交通信息化;2014年01期

2 吳新宇;郭會(huì)文;李楠楠;王歡;陳彥倫;;基于視頻的人群異常事件檢測(cè)綜述[J];電子測(cè)量與儀器學(xué)報(bào);2014年06期

3 薄璐;薄瑋;;智能交通異常事件檢測(cè)系統(tǒng)的分析與設(shè)計(jì)[J];微型電腦應(yīng)用;2014年06期

4 劉振娟;周培建;李宏光;林曉琳;;基于一類賦時(shí)模糊Petri網(wǎng)的化工過程異常事件監(jiān)控[J];化工學(xué)報(bào);2008年07期

5 李兵;李秀;劉文煌;鄭啟鵬;;基于異常感知的動(dòng)態(tài)維修服務(wù)管理[J];計(jì)算機(jī)集成制造系統(tǒng);2006年08期

6 鄭誠;舒堅(jiān);;多尺度時(shí)間序列異常事件檢測(cè)[J];計(jì)算機(jī)工程與應(yīng)用;2006年31期

相關(guān)會(huì)議論文 前1條

1 申麗萍;姜玉杰;隋榮昌;紀(jì)守新;王建生;;開展QC小組活動(dòng),縮短隧道異常事件處理時(shí)間[A];公路交通與建設(shè)論壇（2011）[C];2003年

相關(guān)碩士學(xué)位論文 前10條

1 肖華強(qiáng);基于達(dá)芬奇平臺(tái)的視頻異常事件檢測(cè)算法研究與實(shí)現(xiàn)[D];電子科技大學(xué);2015年

2 余昊;基于底層特征的視頻異常事件檢測(cè)算法研究與實(shí)現(xiàn)[D];上海交通大學(xué);2015年

3 馬金秀;基于視頻的無監(jiān)督異常事件檢測(cè)[D];電子科技大學(xué);2016年

4 彭芊;基于在線自適應(yīng)字典學(xué)習(xí)的視頻異常事件檢測(cè)技術(shù)研究[D];哈爾濱工業(yè)大學(xué);2015年

5 周靈娟;視頻中異常事件檢測(cè)與特征稀疏表示研究[D];杭州電子科技大學(xué);2016年

6 洪嘉鳴;基于數(shù)據(jù)的城市供水管網(wǎng)建模分析和異常事件偵測(cè)[D];杭州電子科技大學(xué);2016年

7 劉威;視頻異常事件檢測(cè)算法設(shè)計(jì)與實(shí)現(xiàn)[D];江西財(cái)經(jīng)大學(xué);2016年

8 劉順華;基于監(jiān)控視頻的人群異常事件檢測(cè)[D];鄭州大學(xué);2017年

9 羅丹;臨床護(hù)理人員對(duì)臨床異常事件報(bào)告現(xiàn)狀及對(duì)策研究[D];第二軍醫(yī)大學(xué);2010年

10 彭寅艷;高速公路隧道條件下異常事件影響的模擬研究[D];重慶大學(xué);2015年

，

本文編號(hào)：1906705