【摘要】：隨著人們安防意識的提高以及數(shù)字視頻技術(shù)的提升,視頻監(jiān)控網(wǎng)絡(luò)已成為維護(hù)社會治安的一個重要組成部分。視頻濃縮系統(tǒng)能對長時間、大范圍的監(jiān)控視頻進(jìn)行濃縮摘要,提高信息挖掘的效率。在視頻監(jiān)控網(wǎng)絡(luò)中常常面臨對指定行人目標(biāo)進(jìn)行識別分析的需求,因此在視頻濃縮系統(tǒng)中實(shí)現(xiàn)行人再辨識有著重要的現(xiàn)實(shí)意義。本文結(jié)合視頻濃縮系統(tǒng)相關(guān)技術(shù),對現(xiàn)有行人再辨識技術(shù)做出了一些改進(jìn),主要工作如下:第一,提出了一種無監(jiān)督的實(shí)時行人再辨識方法。為應(yīng)對海量監(jiān)控視頻行人目標(biāo)圖像,本文提出了一個兩級搜索的框架:在一級搜索建庫時,針對每個行人目標(biāo)圖像提取快速穩(wěn)健的顏色、紋理等局部顯著特征并進(jìn)行降維處理。在檢索時,首先,根據(jù)給定的一張查詢行人目標(biāo)圖片,在數(shù)據(jù)庫中進(jìn)行快速線性相似性度量,篩選出容量較小的結(jié)果候選集;然后,二級搜索時再在候選集中對每張行人目標(biāo)圖像提取相對耗時的局部特征進(jìn)行二次相似度匹配,重新排序得到最后的再辨識結(jié)果。為加快匹配速度,本文設(shè)計(jì)了一種提取速度較高、區(qū)分性較好的局部描述子并進(jìn)行VLAD編碼(MLD-VLAD)。通過相應(yīng)對比實(shí)驗(yàn)表明,在行人再辨識技術(shù)中,與顏色空間相融合的MLD-VLAD取得比SIFT局部描述子更好的區(qū)分性能。第二,提出了一種有監(jiān)督的實(shí)時行人再辨識方法。在無監(jiān)督實(shí)時行人再辨識的基礎(chǔ)上引入了大間隔近鄰(LMNN)算法,并做出了相應(yīng)改進(jìn)。本文提出了基于皮爾遜相關(guān)距離的大間隔近鄰算法(p-LMNN),旨在訓(xùn)練一個線性轉(zhuǎn)換矩陣L,利用L對原始特征以降維的方式進(jìn)行投影,在新的特征空間內(nèi)進(jìn)行距離度量。通過實(shí)驗(yàn)表明,p-LMNN要取得比LMNN更好的再辨識性能表現(xiàn)。第三,在實(shí)時行人再辨識的基礎(chǔ)上,擴(kuò)展出了實(shí)時汽車再辨識。由于汽車本身特征較為剛性,即形狀、姿態(tài)變化較小,無監(jiān)督的再辨識模式已經(jīng)取得較好的表現(xiàn)。實(shí)時汽車再辨識模式有較好的推廣泛化能力,更換應(yīng)用場景無需人工采集樣本進(jìn)行標(biāo)注,簡單易用。
[Abstract]:With the improvement of people's security awareness and the improvement of digital video technology, video surveillance network has become an important part of maintaining public order. Video concentration system can concentrate the long-time and large-scale surveillance video and improve the efficiency of information mining. In video surveillance networks, it is often faced with the need to identify and analyze designated pedestrian targets, so it is of great practical significance to realize pedestrian re-identification in video concentration systems. In this paper, some improvements have been made to the existing pedestrian reidentification technology in combination with the video concentration system. The main work is as follows: first, an unsupervised real-time pedestrian reidentification method is proposed. In order to deal with the massive surveillance video pedestrian target images, this paper proposes a two-level search framework. In the first level searching and building database, the fast and robust local salient features such as color and texture are extracted and dimensionally reduced for each pedestrian target image. In the retrieval, first, according to a given query pedestrian target image, a fast linear similarity measure is carried out in the database, and the candidate set with small capacity is screened out. In the second level search, the local features of each pedestrian target image are extracted by quadratic similarity matching in the candidate set, and the final recognition result is obtained by reordering. In order to speed up the matching speed, this paper designs a local descriptor with high extraction speed and good discriminant and carries out VLAD coding (MLD-VLAD). The corresponding experiments show that the MLD-VLAD fusion with color space is better than the SIFT local descriptor in pedestrian reidentification. Secondly, a supervised real-time pedestrian re-identification method is proposed. Based on the unsupervised real-time pedestrian reidentification, a large interval nearest neighbor (LMNN) algorithm is introduced and improved accordingly. In this paper, a large interval nearest neighbor algorithm (p-LMNN) based on Pearson correlation distance is proposed, which aims at training a linear transformation matrix L, projecting the original feature in a dimensionality reduction way using L, and measuring the distance in a new feature space. The experimental results show that p-LMNN is better than LMNN in re-identification performance. Thirdly, on the basis of real-time pedestrian recognition, real-time vehicle re-identification is extended. Due to the relatively rigid characteristics of the vehicle, that is, the shape and attitude change is small, the unsupervised re-identification model has achieved better performance. The real-time vehicle reidentification model has better generalization ability, and it is simple and easy to use for changing the application scene without manually collecting samples for labeling.
【學(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.41

【相似文獻(xiàn)】

相關(guān)期刊論文 前3條

1 李紹華;馮晶瑩;樓偶俊;景雨;;基于Contourlet變換和神經(jīng)網(wǎng)絡(luò)的視頻水印算法[J];電腦知識與技術(shù);2014年11期

2 張慶濤;;網(wǎng)絡(luò)視音頻系統(tǒng)前臺界面設(shè)計(jì)與實(shí)現(xiàn)[J];電腦知識與技術(shù);2014年11期

3 ;Datavideo TV-1000HD/SD虛擬視頻摳像系統(tǒng)[J];影視制作;2014年04期

相關(guān)重要報(bào)紙文章 前1條

1 吳光兵　王定國;甕福磷酸濃縮系統(tǒng)提升運(yùn)行率[N];中國化工報(bào);2009年

相關(guān)博士學(xué)位論文 前2條

1 伍博;基于顯著性的視覺目標(biāo)跟蹤研究[D];電子科技大學(xué);2017年

2 楊超宇;基于計(jì)算機(jī)視覺的目標(biāo)檢測跟蹤及特征分類研究[D];中國礦業(yè)大學(xué)(北京);2017年

相關(guān)碩士學(xué)位論文 前10條

1 李鵬;視頻濃縮系統(tǒng)中的行人目標(biāo)再辨識技術(shù)[D];北京郵電大學(xué);2016年

2 何慶強(qiáng);基于深度學(xué)習(xí)的視頻內(nèi)容識別技術(shù)研究[D];電子科技大學(xué);2017年

3 李志榮;“喜”“悲”情志激活標(biāo)準(zhǔn)化視頻材料庫的建立[D];山東中醫(yī)藥大學(xué);2017年

4 徐磊;視頻事件結(jié)構(gòu)化描述方法研究[D];南京郵電大學(xué);2017年

5 王睿;網(wǎng)絡(luò)視頻平臺運(yùn)營模式分析[D];北京印刷學(xué)院;2017年

6 王領(lǐng);基于內(nèi)容的視頻拷貝檢測[D];大連理工大學(xué);2017年

7 宿青;基于外部指數(shù)特征的網(wǎng)絡(luò)短視頻推薦方法研究[D];山東科技大學(xué);2017年

8 汪恭焰;基于對象的動態(tài)視頻濃縮和檢索技術(shù)研究[D];長春理工大學(xué);2017年

9 楊澤忠;微信小視頻中的人臉識別關(guān)鍵技術(shù)研究[D];海南大學(xué);2017年

10 陳東浩;視頻走秀場景下的服裝匹配與推薦[D];北京郵電大學(xué);2017年

，

本文編號：2151275