【摘要】：智慧城市是國家解決當(dāng)前城市發(fā)展問題、增加新的經(jīng)濟(jì)增長點(diǎn)、搶占未來科技制高點(diǎn)的重要戰(zhàn)略,其核心建設(shè)內(nèi)容之一是智能交通。智能交通的關(guān)鍵技術(shù)大多涉及計(jì)算機(jī)視覺。本文利用空時(shí)關(guān)系學(xué)習(xí)對復(fù)雜場景下計(jì)算機(jī)視覺中運(yùn)動(dòng)目標(biāo)檢測和目標(biāo)跟蹤兩個(gè)核心問題進(jìn)行了技術(shù)探索,研究成果應(yīng)用于智能交通之智能電子警察系統(tǒng),提高了電子警察系統(tǒng)對環(huán)境的適應(yīng)性。對于運(yùn)動(dòng)目標(biāo)檢測問題,分析了面對復(fù)雜場景代表性的運(yùn)動(dòng)檢測方法設(shè)計(jì)中存在的不足,歸納出形成復(fù)雜場景的主要因素,深入分析了光照變化、背景擾動(dòng)、相似目標(biāo)、相機(jī)運(yùn)動(dòng)等因素對運(yùn)動(dòng)目標(biāo)檢測產(chǎn)生不利影響的機(jī)理,分別提出了綜合利用視頻圖像序列在不同層面的多個(gè)因素、利用目標(biāo)局部特征和空時(shí)關(guān)系以及利用目標(biāo)與周圍環(huán)境的空時(shí)置信關(guān)系等進(jìn)行運(yùn)動(dòng)目標(biāo)檢測的方法；本文還對未知目標(biāo)的長時(shí)間跟蹤問題進(jìn)行了研究。復(fù)雜場景下的未知目標(biāo)長時(shí)跟蹤面臨的問題包括：目標(biāo)遮擋、目標(biāo)外觀變化、目標(biāo)尺度變化以及目標(biāo)的短暫消失。深入分析了目標(biāo)遮擋以及目標(biāo)外觀變化等情況造成目標(biāo)特征缺失或者不完整的情況下,仍可利用的信息,分析并比較了代表性目標(biāo)跟蹤算法應(yīng)對目標(biāo)尺度變化和目標(biāo)短暫消失的處理策略,提出了一種結(jié)合目標(biāo)自身特征和目標(biāo)與周圍環(huán)境的空時(shí)聯(lián)系,可以長時(shí)間對未知目標(biāo)進(jìn)行穩(wěn)定跟蹤的方法；最后,將以上研究成果應(yīng)用于智能電子警察系統(tǒng),解決了研發(fā)過程中遇到的技術(shù)困難。本文的主要研究成果和貢獻(xiàn)：1.分析了視頻目標(biāo)檢測中復(fù)雜場景的主要組成因素,提出一種基于尺度不變局部三元模式(SILTP)的視頻圖像背景建模算法。根據(jù)復(fù)雜場景對視頻圖像序列不同層次的不同影響,利用圖像幀級、圖像塊級和像素級三級信息設(shè)計(jì)背景建模算法。算法融合圖像幀、圖像塊和圖像像素三個(gè)層面的優(yōu)勢來應(yīng)對復(fù)雜場景。在圖像幀級,利用全局灰度均值處理場景亮度突變；在圖像塊級,利用SILTP紋理圖像基于圖像塊進(jìn)行背景建模,快速定位前景目標(biāo)大致輪廓;在像素級,用類ViBe算法檢測前景目標(biāo)精確邊界。此算法挖掘空時(shí)信息并融合利用,其性能在標(biāo)準(zhǔn)視頻集CDM’14上得到驗(yàn)證。2.面對視頻目標(biāo)檢測的難點(diǎn)一目標(biāo)自身投影的消除問題,構(gòu)建了陰影光照模型,分析了目標(biāo)陰影的種類及產(chǎn)生的原因。將紋理信息、色調(diào)信息和空時(shí)信息與ViBe算法相結(jié)合,提出了SAViBe+算法。首先,利用圖像紋理對光照變化的弱敏感性,消除室內(nèi)弱光照產(chǎn)生的目標(biāo)投影；然后,在HSV顏色空間構(gòu)建色調(diào)(Hue)模型,利用物體顏色的固有特性消除室外光照造成的目標(biāo)投影；最后,為了加強(qiáng)目標(biāo)投影的消除效果,同時(shí)提高處理速度,利用像素變化的局部相關(guān)性設(shè)計(jì)了MofV因子。用標(biāo)準(zhǔn)視頻集CDM’14驗(yàn)證了該算法的性能。3.提出在HSV顏色空間實(shí)現(xiàn)魯棒運(yùn)動(dòng)檢測的方法DMSTAB。在HSV顏色空間,通過K-means聚類,利用像素集的空時(shí)關(guān)聯(lián)產(chǎn)生像素的局部強(qiáng)度差,利用單高斯模型分別為像素的局部強(qiáng)度差和色調(diào)建模,然后,聯(lián)合兩者的結(jié)果尋找潛在的陰影像素點(diǎn)；接著,深入分析了ViBe背景差算法的工作原理,提出基于AdaBoost-Like方法利用潛在的陰影像素點(diǎn)構(gòu)建雙關(guān)聯(lián)背景模型,實(shí)現(xiàn)對運(yùn)動(dòng)目標(biāo)快速精確的檢測,有效消除運(yùn)動(dòng)目標(biāo)的自身投影。用標(biāo)準(zhǔn)視頻集CDM’14上多種復(fù)雜場景驗(yàn)證了該方法的性能。4.提出基于空時(shí)置信關(guān)系進(jìn)行運(yùn)動(dòng)目標(biāo)檢測的方法STR。本文提出一種空時(shí)置信關(guān)系,定義了像素點(diǎn)與其環(huán)境鄰域像素點(diǎn)之間一種相對穩(wěn)定的聯(lián)系。首先,根據(jù)視覺聚焦特性和光照影響圖像亮度變化的規(guī)律,定義像素點(diǎn)與環(huán)境像素點(diǎn)的空域關(guān)系；然后,利用快速核密度估計(jì)方法對空域關(guān)系的時(shí)域變化建模；此外,根據(jù)空域關(guān)系值的分散度為模型分配相應(yīng)的權(quán)重；最后,通過基于權(quán)重的概率綜合得到像素點(diǎn)屬于背景的概率,完成運(yùn)動(dòng)目標(biāo)檢測。該算法性能在標(biāo)準(zhǔn)視頻集CDM’14的典型復(fù)雜場景中得到驗(yàn)證。5.提出一種將目標(biāo)與其環(huán)境的空時(shí)關(guān)聯(lián)信息和目標(biāo)自身特征結(jié)合使用,對未知目標(biāo)進(jìn)行長時(shí)間、穩(wěn)定跟蹤的新方法LST。該方法借鑒TLD算法框架,通過檢測和跟蹤兩種獨(dú)立途徑對目標(biāo)進(jìn)行跟蹤。算法包括檢測、跟蹤和學(xué)習(xí)三個(gè)功能模塊。檢測模塊通過若干分類器級聯(lián),根據(jù)目標(biāo)自身基本的圖像特征在全局范圍內(nèi)檢測目標(biāo),處理目標(biāo)短暫消失又重現(xiàn)、目標(biāo)尺度變化以及環(huán)境干擾；跟蹤模塊利用目標(biāo)與其周圍環(huán)境的空時(shí)置信關(guān)系,通過局部搜索,快速跟蹤目標(biāo),處理目標(biāo)遮擋、目標(biāo)尺度變化；算法在運(yùn)行過程中,通過維護(hù)一組由正樣本組成的在線模板,對跟蹤和檢測效果進(jìn)行評測。學(xué)習(xí)模塊依據(jù)評測結(jié)果,調(diào)整檢測模塊和跟蹤模塊相關(guān)參數(shù),實(shí)現(xiàn)算法的自學(xué)習(xí)。在若干對跟蹤算法極具挑戰(zhàn)性(嚴(yán)重遮擋、劇烈的光照變化、姿態(tài)和尺度變化、非剛性形變、復(fù)雜背景、運(yùn)動(dòng)模糊和相似目標(biāo))的數(shù)據(jù)集上比較了LST算法與主流視頻目標(biāo)跟蹤算法的性能,LST算法展現(xiàn)出了較好的跟蹤效果。6.面對電子警察系統(tǒng)研發(fā)過程中遇到的技術(shù)瓶頸,將運(yùn)動(dòng)目標(biāo)檢測算法STR和目標(biāo)跟蹤算法LST的核心技術(shù)應(yīng)用于智能電子警察系統(tǒng)。提高了智能電子警察系統(tǒng)的車輛檢測和車輛跟蹤性能,并進(jìn)一步作用于車牌識(shí)別和車輛違章行為評判,提高了電子警察系統(tǒng)的整體性能。該電子警察系統(tǒng)首期工程已經(jīng)通過驗(yàn)收。
[Abstract]:Intelligent city is an important strategy for a country to solve the current urban development problems, increase new economic growth points, and seize the commanding heights of future science and technology. Intelligent traffic is one of its core construction contents. Two core problems of detection and target tracking are explored technically. The research results are applied to the intelligent electronic police system of intelligent transportation, which improves the adaptability of the electronic police system to the environment. The main factors of complex scenes are analyzed. The mechanism of unfavorable effects of illumination changes, background disturbance, similar targets, camera motion and other factors on moving target detection is analyzed in depth. The problem of long-term tracking of unknown targets in complex scenes includes: object occlusion, object appearance change, object scale change and short-term disappearance of the target. In the case of target feature missing or incomplete due to the change of target appearance, the available information is analyzed and compared. A representative target tracking algorithm is proposed to deal with the change of target scale and the short-term disappearance of target. Finally, the above research results are applied to the intelligent electronic police system to solve the technical difficulties encountered in the development process. The main research results and contributions of this paper are as follows: 1. The main components of complex scenes in video object detection are analyzed, and a scale-invariant approach is proposed. Local ternary mode (SILTP) video image background modeling algorithm. According to the different effects of complex scenes on different levels of video image sequences, the background modeling algorithm is designed using three levels of information: frame level, image block level and pixel level. The algorithm combines the advantages of image frame, image block and image pixel to deal with complex scenes. At the frame level, the global gray mean is used to deal with the sudden change of scene brightness; at the image block level, the SILTP texture image is used to model the background based on the image block to quickly locate the outline of the foreground target; at the pixel level, the precise boundary of the foreground target is detected by the ViBe-like algorithm. Confronted with the difficulty of video object detection, i.e. the elimination of object self-projection, a shadow illumination model is constructed, and the types and causes of object shadows are analyzed. The weak sensitivity of illumination changes eliminates the target projection caused by indoor weak illumination; then, a hue model is constructed in HSV color space to eliminate the target projection caused by outdoor illumination by using the intrinsic characteristics of object color; finally, in order to enhance the elimination effect of target projection and improve the processing speed, the local correlation of pixel changes is used. MofV factor is designed. The performance of the algorithm is verified by the standard video set CDM'14. 3. A robust motion detection method DMSTAB is proposed in HSV color space. In HSV color space, the local intensity difference of pixels is generated by K-means clustering, and the local intensity difference of pixels is generated by spatial-temporal correlation of pixel sets. Then, the working principle of Vibe background subtraction algorithm is deeply analyzed, and a bi-correlation background model based on AdaBoost-Like method is proposed to detect moving objects quickly and accurately and eliminate moving objects effectively. Projection. The performance of this method is validated by a variety of complex scenes on the standard video set CDM'14. 4. A space-time confidence relation based moving object detection method STR is proposed. In this paper, a space-time confidence relation is proposed, and a relatively stable relation between pixels and their neighborhood pixels is defined. Then, a fast kernel density estimation method is used to model the temporal variation of spatial relationship. In addition, the corresponding weights are assigned to the model according to the dispersion of spatial relationship values. Finally, the pixels are synthesized by the probability based on weights. The algorithm is validated in typical complex scenes of standard video set CDM'14. 5. A new method LST is proposed, which combines the space-time association information of the target and its environment with the target's own characteristics to track unknown targets for a long time and stably. The algorithm consists of three functional modules: detection, tracking and learning. The detection module cascades through several classifiers, detects the target in the global scope according to the basic image features of the target itself, handles the transient disappearance and recurrence of the target, changes in target scale and environment. The tracking module uses the space-time confidence relationship between the target and its surroundings to track the target quickly through local search, deal with the occlusion of the target and the change of the target scale; the algorithm evaluates the tracking and detection effect by maintaining a set of online templates composed of positive samples in the running process. The learning module adjusts the tracking and detection results according to the evaluation results. The LST algorithm is compared with the mainstream video target tracking algorithm on several datasets which are challenging to the tracking algorithm (severe occlusion, drastic illumination changes, attitude and scale changes, non-rigid deformation, complex background, motion blur and similar targets). ST algorithm shows a good tracking effect. 6. In the face of the technical bottleneck encountered in the development of the electronic police system, the core technology of moving target detection algorithm STR and target tracking algorithm LST is applied to the intelligent electronic police system. The vehicle detection and tracking performance of the intelligent electronic police system are improved, and further acts on it. License plate recognition and vehicle violation judgment have improved the overall performance of the electronic police system.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級別】：博士
【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10條

1 張濤;張桂林;;運(yùn)動(dòng)檢測中的背景建立和更新[J];計(jì)算機(jī)與數(shù)字工程;2007年02期

2 劉振宇;周莉;陳杰;;一種基于運(yùn)動(dòng)檢測的碼率控制算法[J];科學(xué)技術(shù)與工程;2010年20期

3 閆石;白振興;代忠;;對傳統(tǒng)運(yùn)動(dòng)檢測算法改進(jìn)的研究[J];現(xiàn)代電子技術(shù);2007年06期

4 顧敏劍;;基于運(yùn)動(dòng)檢測的遠(yuǎn)程圖像采集系統(tǒng)的設(shè)計(jì)與實(shí)現(xiàn)[J];計(jì)算機(jī)與數(shù)字工程;2012年08期

5 強(qiáng)俊;周鳴爭;;基于正交Gaussian-Hermite矩的運(yùn)動(dòng)檢測在去隔行中的研究[J];安徽工程大學(xué)學(xué)報(bào);2013年02期

6 胡俊,蘇祥芳,劉立海,沈蕓,管鮑,王延平;圖像序列運(yùn)動(dòng)檢測算法的研究及其應(yīng)用[J];武漢大學(xué)學(xué)報(bào)(自然科學(xué)版);2000年05期

7 周西漢,劉勃,周荷琴;一種基于對稱差分和背景消減的運(yùn)動(dòng)檢測方法[J];計(jì)算機(jī)仿真;2005年04期

8 藍(lán)照華;傅文利;趙進(jìn)創(chuàng);陳濤;;邊緣面積值絕對差數(shù)累積運(yùn)動(dòng)檢測算法[J];微計(jì)算機(jī)信息;2006年33期

9 張凱;;視頻運(yùn)動(dòng)檢測算法的研究和分析[J];遼寧工學(xué)院學(xué)報(bào);2007年01期

10 侯葉;郭寶龍;;基于圖切割的人體運(yùn)動(dòng)檢測[J];光電子.激光;2007年06期

相關(guān)會(huì)議論文 前8條

1 郭銳;;運(yùn)動(dòng)檢測報(bào)警在監(jiān)控系統(tǒng)中的設(shè)計(jì)與實(shí)現(xiàn)[A];科學(xué)發(fā)展與社會(huì)責(zé)任（A卷）——第五屆沈陽科學(xué)學(xué)術(shù)年會(huì)文集[C];2008年

2 唐曉丹;苗振江;;視頻監(jiān)控系統(tǒng)中人的運(yùn)動(dòng)檢測方法研究[A];第十三屆全國信號處理學(xué)術(shù)年會(huì)（CCSP-2007）論文集[C];2007年

3 李東光;殷俊;房慧敏;;基于生物復(fù)眼結(jié)構(gòu)的視覺運(yùn)動(dòng)檢測研究[A];2004全國光學(xué)與光電子學(xué)學(xué)術(shù)研討會(huì)、2005全國光學(xué)與光電子學(xué)學(xué)術(shù)研討會(huì)、廣西光學(xué)學(xué)會(huì)成立20周年年會(huì)論文集[C];2005年

4 傅松寅;蔣剛毅;郁梅;;一種基于像素變化檢測的自適應(yīng)實(shí)時(shí)運(yùn)動(dòng)檢測系統(tǒng)[A];第18屆全國多媒體學(xué)術(shù)會(huì)議（NCMT2009）、第5屆全國人機(jī)交互學(xué)術(shù)會(huì)議（CHCI2009）、第5屆全國普適計(jì)算學(xué)術(shù)會(huì)議（PCC2009）論文集[C];2009年

5 陳威宇;孟利民;;視頻圖像序列混合運(yùn)動(dòng)檢測方法[A];2008年中國高校通信類院系學(xué)術(shù)研討會(huì)論文集（上冊）[C];2009年

6 徐蕭蕭;陳宗海;;基于視覺信息的目標(biāo)檢測與跟蹤技術(shù)現(xiàn)狀與趨勢[A];2007系統(tǒng)仿真技術(shù)及其應(yīng)用學(xué)術(shù)會(huì)議論文集[C];2007年

7 馬強(qiáng);羅喜伶;;空基交通監(jiān)視系統(tǒng)中的運(yùn)動(dòng)目標(biāo)檢測方法研究[A];2008第四屆中國智能交通年會(huì)論文集[C];2008年

8 辛穎;吳強(qiáng);孫光民;徐穎;姚明;;智能化公交系統(tǒng)的客流跟蹤計(jì)數(shù)綜述[A];2008通信理論與技術(shù)新進(jìn)展——第十三屆全國青年通信學(xué)術(shù)會(huì)議論文集（上）[C];2008年

相關(guān)博士學(xué)位論文 前4條

1 范志輝;基于空時(shí)關(guān)系學(xué)習(xí)的運(yùn)動(dòng)檢測和目標(biāo)跟蹤研究[D];西安電子科技大學(xué);2016年

2 李宏友;基于視頻的目標(biāo)檢測與跟蹤方法研究[D];重慶大學(xué);2009年

3 沈宇鍵;變參數(shù)圖像回歸處理方法的研究[D];中國科學(xué)院長春光學(xué)精密機(jī)械與物理研究所;2000年

4 康鋒;基于視覺特征的早期農(nóng)林火災(zāi)檢測方法的基礎(chǔ)研究[D];浙江大學(xué);2010年

相關(guān)碩士學(xué)位論文 前10條

1 李運(yùn)崇;基于形狀識(shí)別的運(yùn)動(dòng)物體檢測[D];鄭州大學(xué);2015年

2 樊建霞;家庭環(huán)境下的人體跟蹤與定位[D];山東大學(xué);2015年

3 余傳桂;網(wǎng)絡(luò)化數(shù)字健身系統(tǒng)設(shè)計(jì)與實(shí)現(xiàn)[D];南昌大學(xué);2015年

4 吳衛(wèi)東;基于視覺注意機(jī)制的視頻顯著目標(biāo)檢測技術(shù)研究[D];北京工業(yè)大學(xué);2015年

5 周文彬;去隔行算法的FPGA實(shí)現(xiàn)[D];西安電子科技大學(xué);2014年

6 喬鵬;基于視頻的車輛運(yùn)動(dòng)檢測和流量統(tǒng)計(jì)算法研究[D];國防科學(xué)技術(shù)大學(xué);2013年

7 李冰冰;面向?qū)W生人群的運(yùn)動(dòng)檢測算法研究及軟件開發(fā)[D];浙江工業(yè)大學(xué);2015年

8 衛(wèi)偉;基于FPGA的2D轉(zhuǎn)3D實(shí)時(shí)視頻轉(zhuǎn)換技術(shù)的研究及實(shí)現(xiàn)[D];合肥工業(yè)大學(xué);2015年

9 李飛;智能視頻監(jiān)控系統(tǒng)中運(yùn)動(dòng)檢測算法的研究[D];重慶大學(xué);2015年

10 白金輝;面向能量回收的MIMU人體運(yùn)動(dòng)檢測識(shí)別方法研究及系統(tǒng)實(shí)現(xiàn)[D];東南大學(xué);2015年

，

本文編號：2221594