發(fā)布時間：2018-04-01 06:57

  本文選題：粒子濾波　切入點：移動目標(biāo)　出處：《新疆大學(xué)》2017年碩士論文

【摘要】：視頻監(jiān)控對目標(biāo)展開跟蹤的計算是近些年來在計算機(jī)領(lǐng)域中應(yīng)用非常廣泛的熱點項目之一。該項技術(shù)不論是在軍事領(lǐng)域上的應(yīng)用還是在民用領(lǐng)域上的應(yīng)用,都有著非常重要的研究價值。隨著科技的進(jìn)步,對粒子濾波的使用和研究開始得到了越來越多領(lǐng)域的重視。并且從近些年來看,粒子濾波算法已經(jīng)被廣泛地應(yīng)用到了視頻監(jiān)控中,對要追蹤的目標(biāo)進(jìn)行跟蹤。在視頻監(jiān)控中對移動目標(biāo)進(jìn)行監(jiān)控的過程是非常復(fù)雜的,其中不僅包括了目標(biāo)物體的動態(tài)變化,還包括了移動目標(biāo)之間可能發(fā)生的遮擋、合并以及分離等情況。因此,本文的研究目的與意義旨在即粒子濾波算法的技術(shù)應(yīng)用之上,加強(qiáng)對視頻中移動目標(biāo)的出現(xiàn)進(jìn)行應(yīng)有的跟蹤和必要的分析,解決好目標(biāo)在移動過程中可能出現(xiàn)的復(fù)雜的情形,然后利用好粒子濾波的計算方法,從而實現(xiàn)對移動目標(biāo)在視頻中被成功跟蹤及進(jìn)一步追蹤的觀察目的。本文在行文的過程中對視頻中移動的目標(biāo)主要運用了粒子濾波算法與Mean shift算法,并最終對這兩種算法進(jìn)行了跟蹤實驗效果對比。在正文整體結(jié)構(gòu)上:第一,本文先介紹了幾種比較常見的目標(biāo)跟蹤算法,并對這些算法的特點進(jìn)行了相關(guān)分析與說明;第二,本文論述了與粒子濾波相關(guān)的理論框架以及利用粒子濾波對移動目標(biāo)進(jìn)行跟蹤的算法,并且用具體的視頻跟蹤實驗說明了粒子濾波在跟蹤視頻中對移動目標(biāo)進(jìn)行跟蹤的可靠性;第三,介紹Mean shift算法,并且通過對同一個視頻圖像中的移動目標(biāo)進(jìn)行跟蹤實驗,對比發(fā)現(xiàn)這兩種算法的各自特點,為后續(xù)工作做準(zhǔn)備;第四,在先前內(nèi)容所研究的基礎(chǔ)上,總結(jié)出了本文的核心內(nèi)容“基于粒子濾波算法與Mean shift算法結(jié)合使用基礎(chǔ)之上”的對視頻中移動目標(biāo)進(jìn)行跟蹤和計算的算法。這個算法可以有效的對被跟蹤的目標(biāo)在遮擋發(fā)生之后再次對被跟蹤目標(biāo)進(jìn)行跟蹤處理,能在一定程度上對移動目標(biāo)在發(fā)生遮擋時容易造成跟蹤丟失的現(xiàn)象起到彌補(bǔ)作用,同時該算法也能夠在最快的時間內(nèi)鑒別出移動目標(biāo)可能發(fā)生的情況,從而進(jìn)行結(jié)合算法的選擇性應(yīng)用,繼而來應(yīng)對移動目標(biāo)被遮擋的現(xiàn)象,并且還可以有效地降低視頻噪聲及雜波的干擾,從而有效地提高視頻跟蹤系統(tǒng)對移動目標(biāo)的跟蹤。
[Abstract]:The computation of video surveillance and target tracking is one of the most popular projects in the computer field in recent years. This technology is used in both military and civilian fields. With the development of science and technology, the use and research of particle filter have been paid more and more attention. And in recent years, Particle filter algorithm has been widely used in video surveillance, tracking the target to be tracked. In video surveillance, the process of monitoring moving target is very complicated, which includes not only the dynamic changes of the target object, but also the dynamic change of the target object. It also includes the possible occlusion, merging and separation between moving targets. Therefore, the purpose and significance of this paper is to apply the particle filter algorithm. We should track and analyze the moving targets in video, solve the complex situations that may occur in the moving process, and then make good use of the particle filter calculation method. In this paper, particle filter algorithm and Mean shift algorithm are mainly used to track moving target in video. Finally, the experimental results of the two algorithms are compared. In the overall structure of the text: first, this paper introduces several common target tracking algorithms, and the characteristics of these algorithms are analyzed and explained. In this paper, the theoretical framework related to particle filter and the algorithm of moving target tracking by particle filter are discussed, and the reliability of particle filter in tracking moving target in tracking video is illustrated by specific video tracking experiment. Thirdly, the Mean shift algorithm is introduced, and the experiment of moving target tracking in the same video image is carried out, and the characteristics of the two algorithms are compared to prepare for the follow-up work. Fourth, on the basis of the previous research, This paper summarizes the algorithm of tracking and calculating moving target in video, which is based on particle filter algorithm and Mean shift algorithm. This algorithm can effectively track the target to be tracked. Track the tracked target again after the occlusion has occurred, To a certain extent, it can make up for the phenomenon that the moving target is likely to lose track when occlusion occurs. At the same time, the algorithm can also identify the possible situation of moving target in the fastest time. Therefore, the selective application of the combined algorithm can deal with the occlusion phenomenon of moving target, and it can also effectively reduce the noise and clutter interference of video, thus effectively improve the tracking of moving target in video tracking system.
【學(xué)位授予單位】：新疆大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 Yan Zhang;Shafei Wang;Jicheng Li;;Improved particle filtering techniques based on generalized interactive genetic algorithm[J];Journal of Systems Engineering and Electronics;2016年01期

2 石雪軍;紀(jì)志成;;基于改進(jìn)粒子濾波的射頻識別室內(nèi)跟蹤算法[J];計算機(jī)工程;2015年11期

3 林h，

本文編號：1694440