【摘要】：近年來(lái),隨著人們自身公共安全意識(shí)的提高以及視頻監(jiān)控技術(shù)的發(fā)展,智能視頻監(jiān)控系統(tǒng)得到了大量的普及。行人再識(shí)別(Person Re-identification)是近幾年智能視頻分析領(lǐng)域興起的一項(xiàng)新技術(shù),是多攝像機(jī)聯(lián)合智能視頻監(jiān)控系統(tǒng)中需要解決的關(guān)鍵問(wèn)題之一,因而得到了廣大計(jì)算機(jī)視覺(jué)領(lǐng)域及人工智能領(lǐng)域相關(guān)人員的關(guān)注。無(wú)重疊視域下的行人再識(shí)別是指在一個(gè)多攝像機(jī)聯(lián)合的視頻監(jiān)控系統(tǒng)下,判斷一個(gè)攝像頭中出現(xiàn)的行人目標(biāo)是否在另一個(gè)非重疊視域監(jiān)控下的攝像頭中出現(xiàn)過(guò)的一個(gè)過(guò)程,即識(shí)別出不同攝像機(jī)拍攝到的屬于某一個(gè)行人的圖像。但由于受攝像機(jī)角度、背景變化、光照條件、姿態(tài)變化、遮擋等多種外在復(fù)雜因素的影響,同一行人在不同視域下可能存在很大的差異性,從而使得行人再識(shí)別問(wèn)題具有很大的挑戰(zhàn)性。本文提出了一種基于度量學(xué)習(xí)的行人再識(shí)別算法PRML(Person Re-identification based on Metric Learning),主要通過(guò)特征學(xué)習(xí)生成一個(gè)測(cè)度矩陣來(lái)進(jìn)行行人的再識(shí)別。本文首先通過(guò)一種圖像增強(qiáng)算法對(duì)原始行人圖像進(jìn)行處理,從而減少因光照變化所帶來(lái)的影響,然后根據(jù)人體目標(biāo)外觀形態(tài)特性對(duì)行人進(jìn)行合理分割,并提取行人圖像顏色特征(HSV、Lab)、紋理特征(SILTP、FHOG)以及顏色屬性ColorNames特征并進(jìn)行核函數(shù)學(xué)習(xí),將原始線性特征空間投影到更加具有區(qū)分性的非線性特征空間并對(duì)特征進(jìn)行PCA降維,之后考慮到不同類(lèi)型特征對(duì)行人圖像描述的差異性,分別學(xué)習(xí)得到三個(gè)獨(dú)立的測(cè)度矩陣,并通過(guò)正則化方法來(lái)優(yōu)化測(cè)度矩陣的過(guò)擬合問(wèn)題,最終并加權(quán)融合多個(gè)測(cè)度矩陣綜合得到行人圖像對(duì)的相似性度量函數(shù),從而實(shí)現(xiàn)行人相似性的度量。最后在VIPeR、iLIDS、CUHK01三個(gè)公共數(shù)據(jù)集上采用CMC(Cumulative Matching Characteristic Curve)曲線評(píng)測(cè)標(biāo)準(zhǔn)對(duì)提出的算法進(jìn)行了實(shí)驗(yàn)效果驗(yàn)證、對(duì)比和分析。
[Abstract]:In recent years, with the improvement of public safety awareness and the development of video surveillance technology, intelligent video surveillance system has been widely used. Pedestrian rerecognition (Person Re-identification) is a new technology emerging in the field of intelligent video analysis in recent years. It is one of the key problems to be solved in multi-camera joint intelligent video surveillance system. As a result, the field of computer vision and artificial intelligence related to the field of attention. Pedestrian reidentification without overlap is a process in which a video surveillance system with multiple cameras is used to determine whether the pedestrian target in one camera has appeared in another camera. That is to identify the different cameras of the image of a pedestrian. However, due to the influence of many complicated external factors, such as camera angle, background change, illumination condition, attitude change, occlusion and so on, the same line of people may have great differences in different visual fields. Therefore, the problem of pedestrian recognition is very challenging. In this paper, a new pedestrian rerecognition algorithm based on metric learning (PRML (Person Re-identification based on Metric Learning),) is proposed, which is based on feature learning to generate a measure matrix for pedestrian rerecognition. In this paper, the original pedestrian image is processed by an image enhancement algorithm, so as to reduce the influence caused by the change of illumination, and then the pedestrian is segmented reasonably according to the appearance and morphological characteristics of the human object. The color feature (HSV), texture feature (SILTPFHOG) and color attribute (ColorNames) of pedestrian image are extracted, and the kernel function is studied. The original linear feature space is projected into a more discriminative nonlinear feature space and the feature dimension is reduced by PCA. Considering the difference of pedestrian image description between different types of features, three independent measure matrices are obtained, and the overfitting problem of measure matrix is optimized by regularization method. Finally, the similarity measurement function of pedestrian image pairs is obtained by combining multiple measure matrices weighted and weighted, and the pedestrian similarity measurement is realized. Finally, the experimental results of the proposed algorithm are verified, compared and analyzed by using CMC (Cumulative Matching Characteristic Curve) curve evaluation standard on three common data sets of VIPeR iLIDSU CUHK01.
【學(xué)位授予單位】：南京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類(lèi)號(hào)】：TP391.41;TN948.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前5條

1 齊美彬;檀勝順;王運(yùn)俠;劉皓;蔣建國(guó);;基于多特征子空間與核學(xué)習(xí)的行人再識(shí)別[J];自動(dòng)化學(xué)報(bào);2016年02期

2 杜宇寧;艾海舟;;基于統(tǒng)計(jì)推斷的行人再識(shí)別算法[J];電子與信息學(xué)報(bào);2014年07期

3 范彩霞;朱虹;藺廣逢;羅磊;;多特征融合的人體目標(biāo)再識(shí)別[J];中國(guó)圖象圖形學(xué)報(bào);2013年06期

4 高勇;;大力推進(jìn)視頻監(jiān)控建設(shè)與應(yīng)用 不斷提升公安工作能力和水平——全國(guó)公安機(jī)關(guān)視頻監(jiān)控建設(shè)與應(yīng)用工作會(huì)議舉行[J];中國(guó)安防;2012年08期

5 張志成;;基于監(jiān)控視頻的分析技術(shù)已成為執(zhí)法調(diào)查的主要手段[J];中國(guó)安防;2012年08期

，

本文編號(hào)：2193585