海量监控视频快速回放与检索技术

为解决海量监控视频的快速浏览和检索,介绍了一种基于目标索引的视频摘要和检索方法。该方法在光流分析的基础上,在画面的静止区域更新背景,运动的区域利用差分法分割出运动目标图像。经过优化的快速特征匹配和建立运动跟踪模型后,根据目标运动轨迹,按照时空距离进行聚类。在目标图像数据和运动参数进行XML结构化存储为索引的基础上,最后在检索时将符合条件的所有目标图像,按照其原有时间顺序逐帧贴到同一个背景图像中,形成动态的摘要视频。由于该方法剔除了背景中大量的时空冗余信息,可在较短回放时间内浏览全部有用目标,显著提高海量监控视频的查阅效率。     

视频监控中一种完整提取运动目标的检测算法

提出一种视频监控中完整、精确提取运动目标前景的检测算法.首先对彩色图像建立混合高斯模型,由背景差分法得到基本准确的前景图像;然后和对称差分法图像综合,得到完整可靠的运动目标图像;再利用亮度信息消除运动目标阴影;最后利用形态学滤波和连通区域面积检测进行后处理.实验结果表明,该算法检测的运动目标前景信息完整准确,对固定场景下的视频监控系统具有一定实用价值.     

基于运动目标分类的监控视频检索系统

设计了一种面向运动目标类别的监控视频检索系统,用于快速有效地进行视频片段的定位。系统由视频分析和视频检索两部分组成,视频分析包含了运动目标检测、跟踪和分类等基本算法模块。基于视频分析模块所提供的运动目标信息,对不同视频片段进行类别标注,并形成目标类别描述文件。在视频检索过程中,使用者可通过输入的运动目标类别,快速查询到相应的视频片段。实验表明,所建系统能够帮助用户在大量的监控视频片段中快速的找出满足指定目标类别的视频片段,避免繁琐的手动浏览和查找,有效提高了监控视频检索的效率。     

AVS监控档视频的压缩域摘要研究

传统的像素域视频摘要方法需要完全解码视频,计算时间长。因此提出了一种在不解码的前提下直接在压缩域生成AVS格式摘要视频的算法。首先对AVS码流中的运动向量(MV)进行分析,提取前景运动宏块;然后跟踪前景宏块,得到有效的运动目标轨迹;最后提取AVS监控档视频中的背景帧,并将它与运动目标轨迹相结合,生成摘要视频。实验表明,该算法能够在压缩域下有效地生成摘要视频,相比于传统方法具有更快的处理速度。     

分布式网络信息数据防篡改方法研究

为在足球视频中有效的检测与跟踪运动目标,需要对足球比赛视频中目标检测与跟踪算法进行研究。当前采用的算法,在动态场景中,存在运动目标检测与跟踪效果不佳的问题。为此,提出一种基于OpenCV的足球比赛视频中目标检测与跟踪算法。该算法结合平均背景算法将足球比赛视频中目标图像分割为前景区与背景区,计算足球比赛视频每一帧目标图像和背景图像之间差值的绝对差值,同时计算每一个目标图像中像素点的平均值与标准值来建立目标图像背景统计模型,利用TMHI算法对足球比赛视频中目标初始图像进行阈值分割,得到初始分割图像,对分割图像进行中值滤波和闭运算,再使用卡尔曼滤波对分割后的目标图像进行处理,得到镜头中目标的质心位置和目标外界矩形框,然后对足球比赛视频中目标进行跟踪。实验证明,该算法有效的检测与跟踪足球视频中运动目标。     

相机运动条件下的行人检测与跟踪

行人检测与跟踪在司机辅助安全系统和视频监控等领域具有重要的地位.针对目前存在的关键问题,如人体运动,相机运动,背景及形状、角度等变化对检测及跟踪带来的干扰,提出了一种将运动信息与形状信息相结合的行人检测方法,准确检测运动摄像机拍摄的直立运动人体;使用了基于小面积目标的跟踪算法进行人体跟踪;利用实际拍摄的视频序列进行算法验证.实验结果表明,混合检测算法速度快,准确率高;基于小面积的跟踪算法能够鲁棒的跟踪检测到的运动人体.     

基于DSP的靶场远程视频监控系统的设计与实现

针对航天靶场的实际情况,设计了一种基于DSP的靶场远程视频监控系统;介绍了系统的总体硬软件结构,详细设计了系统的DSP视频处理模块、网络通信模块、视频接口模块,阐述了视频压缩包的格式和视频数据流的RTP封装方法,提出了一种对视频图像中运动目标的检测和提取算法;通过多次卫星发射任务检验,结果表明,该系统对运动目标的背景噪声具有明显的抑制作用,能够在复杂的图像背景下准确识别运动目标,并对其进行平稳跟踪;该系统对其它远程视频监控系统的设计具有一定的参考价值.     

基于Adaboost和码本模型的手扶电梯出入口视频监控方法

针对传统视频监控方法无法对密集前景目标进行准确分割的问题,提出一种基于Adaboost和码本模型的多目标视频监控方法。首先,通过训练得到Adaboost人头分类器,利用码本算法为垂直拍摄的手扶电梯出入口图像建立背景模型,提取前景图像对其进行人头检测和跟踪;之后,剔除行人目标得到物件目标,对物件目标进行跟踪;最后,根据行人和物件的运动特征进行监控。对12段出入口视频序列的实验结果表明,监控方法能够准确稳定地跟踪行人和物件,完成逆行检测、客流统计、行人拥堵和物件滞留等监控任务,处理速度达到36帧/秒,目标跟踪准确率达到94%以上,行为监控准确率达到95.8%,满足智能视频监控系统鲁棒性、实时性和准确性的要求。     

浅析智能视频监控中的关键技术

智能视频监控技术目前广泛应用于机场、地铁、银行等重要场合的监控系统,因其具有广泛的应用价值,得到国内外学者的关注。该文重点阐述智能视频监控中运动目标检测和运动目标跟踪算法,并对各种算法优缺点进行相关的探讨,对智能视频监控的发展具有一定的意义。     

基于Surendra背景差分和帧间差分的运动目标检测

运动目标检测是视频监控和目标跟踪研究的前提,目标提取是否准确,将直接影响到后期目标跟踪和处理的效果。该文在背景减除法和五帧差分法的基础上,提出了将两种算法相结合的方法,对视频序列中的运动目标进行检测。首先采用Surendra背景减除法建立背景模型,并通过迭代法计算出动态阈值,完成背景更新,然后将检测出的运动目标区域与五帧差分法得到的目标轮廓图形进行逻辑"或"操作,取得比较完整的运动目标,最后采用连通性检测和孔洞填充方法综合得到运动区域图像。通过MATLAB验证和量化评估表明,该方法能准确地检测出运动目标,同时在识别率和误检率方面也得到了较好结果。     

Layered Representation of a Video Shot with Mosaicing

: This paper presents a motion segmentation method useful for representing efficiently a video shot as a static mosaic of the background plus sequences of the objects moving in the foreground. This generates an MPEG-4 compliant, layered representation useful for video coding, editing and indexing. First, a mosaic of the static background is computed by estimating the dominant motion of the scene. This is achieved by tracking features over the video sequence and using a robust technique that discards features attached to the moving objects. The moving objects get removed in the final mosaic by computing the median of the grey levels. Then, segmentation is obtained by taking the pixelwise difference between each frame of the original sequence and the mosaic of the background. To discriminate between the moving object and noise, temporal coherence is exploited by tracking the object in the binarised difference image sequence. The automatic computation of the mosaic and the segmentation procedure are illustrated with real sequences experiments. Examples of coding and content-based manipulation are also shown. Received: 31 August 2000, Received in revised form: 18 April 2001, Accepted: 20 July 2001     

基于视觉传达的后继帧视频图像目标跟踪仿真

现阶段所采用的跟踪方法对后继帧视频图像目标跟踪存在跟踪效果不理想、跟踪效率较低等问题。提出基于视觉传达的后继帧视频图像目标跟踪方法。利用图像差分方法获取视频图像运动目标可能出现的区域,并对这个区域视频图像目标进行运动估计,采用形态学方法来降低聚类区域的数量,得到后继帧视频图像目标区域;采用均值漂移法估计后继帧视图像核概率密度,对后继帧视频图像进行分割处理,找出后继帧视频图像目标区域最显著的特征,通过迭代运算找到目标位置,实现目标跟踪。实验结果表明,所提算法具有较好的后继帧视频图像跟踪效果、并且跟踪效率较高,具有一定的应用价值。     

侦查目标追踪过程中视频浓缩方法仿真

在侦查目标追踪过程中,采用当前方法对视频进行浓缩处理时,所用时间较长,浓缩处理后视频中有效信息丢失率较高。为此,提出新的侦查目标追踪过程中视频浓缩方法。通过混合高斯模型模拟视频信号,实现视频前景建模和背景建模;在Camshift算法下降H分量的颜色概率分布图和颜色直方图相结合,对视频中存在的图像帧进行运算,通过调整搜索窗的大小实现对运动目标的追踪;采用松弛线性规划算法获得目标运动轨迹对应的最优时间标签,结合目标轨迹、背景序列和最优时间标签实现视频的浓缩处理。仿真结果表明,所提方法的浓缩耗时较短,视频中有效信息得到了有效保存。     

Algebraic retrieval of fragmentarily indexed video

When dealing with long video data, the task of identifying and indexing all meaningful subintervals that become answers to some queries is infeasible. It is infeasible not only when done by hand but even when done by using latest automatic video indexing techniques. Whether manually or automatically, it is only fragmentary video intervals that we can identify in advance of any database usage. Our goal is to develop a framework for retrieving meaningful intervals from such fragmentarily indexed video data. We propose a set of algebraic operations that includes ourglue join operations, with which we can dynamically synthesize all the intervals that are conceivably relevant to a given query. In most cases, since these operations also produce irrelevant intervals, we also define variousselection operations that are useful in excluding them from the answer set. We also show the algebraic properties possessed by those operations, which establish the basis of an algebraic query optimization. Katsumi Tanaka, D. Eng.: He received his B.E., M.E., and D.Eng. degrees in information science from Kyoto University, in 1974, 1976, and 1981, respectively. Since 1994, he is a professor of the Department of Computer and Systems Engineering and since 1997, he is a professor of the Division of Information and Media Sciences, Graduate School of Science and Technology, Kobe University. His research interests include object-oriented, multimedia and historical databases abd multimedia information systems. He is a member of the ACM, IEEE Computer Society and the Information Processing Society of Japan. Keishi Tajima, D.Sci.: He received his B.S, M.S., and D.S. from the department of information science of University of Tokyo in 1991, 1993, and 1996 respectively. Since 1996, he is a Research Associate in the Department of Computer and Systems Engineering at Kobe University. His research interests include data models for non-traditional database systems and their query languages. He is a member of ACM, ACM SIGMOD, Information Processing Society of Japan (IPSJ), and Japan Society for Software Science and Technology (JSSST). Takashi Sogo, M.Eng.: He received B.E. and M.E. from the Department of Computer and Systems Engineering, Kobe University in 1998 and 2000, respectively. Currently, he is with USAC Systems Co. His research interests include video database systems. Sujeet Pradhan, D.Eng.: He received his BE in Mechanical Engineering from the University of Rajasthan, India in 1988, MS in Instrumentation Engineering in 1995 and Ph.D. in Intelligence Science in 1999 from Kobe University, Japan. Since 1999 May, he is a lecturer of the Department of Computer Science and Mathematics at Kurashiki University of Science and the Arts, Japan. A JSPS (Japan Society for the Promotion of Science) Research Fellow during the period between 1997 and 1999, his research interests include video databases, multimedia authoring, prototypebased languages and semi-structured databases. Dr. Pradhan is a member of Information Processing Society of Japan.     

Real time multiple objects tracking and identification based on discrete wavelet transform

A new method for detecting and tracking multiple moving objects based on discrete wavelet transform and identifying the moving objects by their color and spatial information is proposed in this paper. Many tracking algorithms have better performance under static background but get worse results under background with fake motions. Therefore, most of the tracking algorithms are used indoors instead of outdoor environment. Since discrete wavelet transform has a nice property that it can divide a frame into four different frequency bands without loss of the spatial information, it is adopted to solve this problem due to the fact that most of the fake motions in the background can be decomposed into the high frequency wavelet sub-band. In tracking multiple moving objects, many applications have problems when objects pass across each other. Color and spatial information are used in this paper to solve this problem. The experimental results prove the feasibility and usefulness of the proposed method.     

基于深度信念网络的高维传感器数据异常检测算法

越来越多的物联网数据呈现高维度特征,针对目前传感器数据异常检测算法对高维数据在线检测的困难,提出一种基于深度信念网络的高维传感器数据异常检测算法。首先利用深度信念网络对高维数据进行特征提取,降低原始数据维度,再对降维后的数据进行异常检测。在检测过程中将QSSVM(Quarter-Sphere Support Vector Machine)与滑动窗口模型相结合,实现了在线式的异常检测。通过在四组真实传感器数据上的大量实验,与先前的异常检测算法做了对比,实验结果表明,新算法相对于OCSVM(One-Class Support Vector Machine)仅利用原有算法50%的计算时间,将检测准确度提高了约20%。     

前景判别的局部模型匹配目标跟踪

目的 在复杂背景下,传统模型匹配的跟踪方法只考虑了目标自身特征,没有充分考虑与其所处图像的关系,尤其是目标发生遮挡时,易发生跟踪漂移,甚至丢失目标。针对上述问题,提出一种前景判别的局部模型匹配(FDLM)跟踪算法。方法 首先选取图像帧序列前m帧进行跟踪训练,将每帧图像分割成若干超像素块。然后,将所有的超像素块组建向量簇,利用判别外观模型建立包含超像素块的目标模型。最后,将建立的目标模型作为匹配模板,采用期望最大化(EM)估计图像的前景信息,通过前景判别进行局部模型匹配,确定跟踪目标。结果 本文算法在前景判别和模型匹配等方面能准确有效地适应视频场景中目标状态的复杂变化,较好地解决各种不确定因素干扰下的跟踪漂移问题,和一些优秀的跟踪算法相比,可以达到相同甚至更高的跟踪精度,在Girl、Lemming、Liquor、Shop、Woman、Bolt、CarDark、David以及Basketball视频序列下的平均中心误差分别为9.76、28.65、19.41、5.22、8.26、7.69、8.13、11.36、7.66,跟踪重叠率分别为0.69、0.61、0.77、0.74、0.80、0.79、0.79、0.75、0.69。结论 实验结果表明,本文算法能够自适应地实时更新噪声模型参数并较准确估计图像的前景信息,排除背景信息干扰,在部分遮挡、目标形变、光照变化、复杂背景等条件下具有跟踪准确、适应性强的特点。     

基于OpenCV的视频运动目标检测与跟踪

针对视频文件中运动目标检测与跟踪这一问题,提出一种先检测后跟踪的方法.首先利用平均背景法完成对背景模型的更新,从而检测出运动目标,在此基础上利用投影法来投影出运动目标的大小,最后再利用MeanShift算法对运动目标进行跟踪.在跟踪过程中,通过OpenCV编写程序实现对运动目标的检测与跟踪.实验验证,该方法在实现运动目标的精确检测与跟踪的基础上,减少了运算量,提高了跟踪的速度.     

Non-interactive automatic video segmentation of moving targets

Extracting moving targets from video accurately is of great significance in the field of intelligent transport.To some extent,it is related to video segmentation or matting.In this paper,we propose a non-interactive automatic segmentation method for extracting moving targets.First,the motion knowledge in video is detected with orthogonal Gaussian-Hermite moments and the Otsu algorithm,and the knowledge is treated as foreground seeds.Second,the background seeds are generated with distance transformation based on foreground seeds.Third,the foreground and background seeds are treated as extra constraints,and then a mask is generated using graph cuts methods or closed-form solutions.Comparison showed that the closed-form solution based on soft segmentation has a better performance and that the extra constraint has a larger impact on the result than other parameters.Experiments demonstrated that the proposed method can effectively extract moving targets from video in real time.     

Object extraction from an image of wear particles on a complex background

The object extraction of a debris image is an important basic task in identifying wear particles in ferrographic analysis. However, there is some difficulty in object extraction because of noise jamming in the original debris image. In the present study, two methods of image enhancement—weighted mean filtering and adaptive median filtering—were applied in order to improve the image quality. Then, the adaptive thresholding selection method was used, which is based on an improved debris image. Finally, the effective segmentation of the debris image and the automatic extraction of debris objects were realized. At the same time, targetting the characteristics of low proportion of an object in the total image, a novel method of adaptive thresholding selection was put forward, which is based on the Ostu thresholding method. The segmentation results along with the debris image prove that the current method can give more precise and accurate segmentation of objects than the classical methods. The results also showed that methods in the present paper were concise and effective, which provides an important basis for the further study of debris recognition, fault diagnosis, and condition monitoring of machines. The text was submitted by the authors in English. Xianguo Hu (born 1963), PhD, is a professor at the School of Mechanical and Automotive Engineering at the Hefei University of Technology, China. He received his BS and MS in Powder Metallurgy Material and Mechanics (Tribology) from the Hefei University of Technology in 1985 and 1988, respectively. His PhD degree was awarded at Szent Istvan University, Hungary, in 2002. As a visiting scientist, he conducted research at the Technical University of Budapest, Hungary, and the Technical University of Berlin, Germany, from 1994 to 1997. His research areas include wear debris analysis, optimal tribological design, friction and wear mechanisms, etc. He is the author or coauthor of more than 100 published technical papers. Peng Huang (born 1981) is an MS student at the School of Mechanical and Automotive Engineering of Hefei University of Technology, China. His main focus is on wear debris analysis. Shousen Zheng (born 1963) is an associate professor at the School of Engineering, SunYat-Sen University, China. He received his BS, MS, and PhD in Mechanical Engineering from Hefei University of Technology in 1985, 1988, and 2001, respectively. From 1988 to 2004, he was employed at the Department of Mechanical Engineering at the Hefei University of Technology. In 2005, he moved to the current university. His research interests include computer language, auto CAD/CAM, wear debris analysis, etc. He is the author or coauthor of more than 40 published technical papers.