本文選題：多目標(biāo)跟蹤 + 識(shí)別輔助的跟蹤　； 參考：《中國(guó)圖象圖形學(xué)報(bào)》2017年03期

【摘要】：目的目標(biāo)的長(zhǎng)距離跟蹤一直是視頻監(jiān)控中最具挑戰(zhàn)性的任務(wù)之一�，F(xiàn)有的目標(biāo)跟蹤方法在存在遮擋、目標(biāo)消失再出現(xiàn)等情況下往往會(huì)丟失目標(biāo),無(wú)法進(jìn)行持續(xù)有效的跟蹤。一方面目標(biāo)消失后再次出現(xiàn)時(shí),將其作為新的目標(biāo)進(jìn)行跟蹤的做法顯然不符合實(shí)際需求;另一方面,在跟蹤過(guò)程中當(dāng)相似的目標(biāo)出現(xiàn)時(shí),也很容易誤導(dǎo)跟蹤器把該相似對(duì)象當(dāng)成跟蹤目標(biāo),從而導(dǎo)致跟蹤失敗。為此,提出一種基于目標(biāo)識(shí)別輔助的跟蹤算法來(lái)解決這個(gè)問(wèn)題。方法將跟蹤問(wèn)題轉(zhuǎn)化為尋找?guī)g檢測(cè)到的目標(biāo)之間對(duì)應(yīng)關(guān)系問(wèn)題,從而在目標(biāo)消失再現(xiàn)后,采用深度學(xué)習(xí)網(wǎng)絡(luò)實(shí)現(xiàn)有效的軌跡恢復(fù),改善長(zhǎng)距離跟蹤效果,并在一定程度上避免相似目標(biāo)的干擾。結(jié)果通過(guò)在標(biāo)準(zhǔn)數(shù)據(jù)集上與同類(lèi)算法進(jìn)行對(duì)比實(shí)驗(yàn),本文算法在目標(biāo)受到遮擋、交叉運(yùn)動(dòng)、消失再現(xiàn)的情況下能夠有效地恢復(fù)其跟蹤軌跡,改善跟蹤效果,從而可以對(duì)多個(gè)目標(biāo)進(jìn)行持續(xù)有效的跟蹤。結(jié)論本文創(chuàng)新性地提出了一種結(jié)合基于深度學(xué)習(xí)的目標(biāo)識(shí)別輔助的跟蹤算法,實(shí)驗(yàn)結(jié)果證明了該方法對(duì)遮擋重現(xiàn)后的目標(biāo)能夠有效的恢復(fù)跟蹤軌跡,適用在監(jiān)控視頻中對(duì)多個(gè)目標(biāo)進(jìn)行持續(xù)跟蹤。
[Abstract]:Target long-range tracking is one of the most challenging tasks in video surveillance. The existing methods of target tracking often lose the target in the presence of occlusion and disappear and reappear, so they can not be tracked continuously and effectively. On the one hand, tracking a target as a new target when it reappears after disappearing is clearly not in line with actual needs; on the other hand, when a similar target appears in the tracking process, It is also easy to mislead the tracker to treat the similar object as a tracking target, resulting in tracking failure. Therefore, a tracking algorithm based on target recognition assistance is proposed to solve this problem. Methods the tracking problem is transformed into the problem of finding the corresponding relationship between the detected targets between frames. After the target vanishes and reappears, the depth learning network is used to achieve effective trajectory recovery and to improve the effect of long distance tracking. And to a certain extent to avoid the interference of similar targets. Results by comparing with the similar algorithms on the standard data set, the algorithm can recover the tracking track effectively and improve the tracking effect when the target is occluded, cross moving, disappear and reappear. Thus, multiple targets can be tracked continuously and effectively. Conclusion in this paper, a target recognition aided tracking algorithm based on depth learning is proposed. The experimental results show that this method can effectively recover the track of the target after occlusion reconstruction. Suitable for continuous tracking of multiple targets in surveillance video.
【作者單位】： 浙江工商大學(xué)計(jì)算機(jī)與信息工程學(xué)院;北京正安維視科技股份有限公司;蘭州大學(xué)信息科學(xué)與工程學(xué)院;
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(61472362,61379075) 浙江省自然科學(xué)基金項(xiàng)目(LZ16F020002,LY14F020001) 公益技術(shù)研究社會(huì)發(fā)展項(xiàng)目(2015C33081)~~
【分類(lèi)號(hào)】：TP391.41
，

本文編號(hào)：1883270