视频监控中基于在线多核学习的目标再现识别

在非重叠多摄像机或单摄像机视频监控中,识别跟踪目标的再次出现很重要.针对传统支持向量机方法在特征融合方面的缺陷,本文提出了一种新的基于在线多核学习的人体目标再现识别方法.该方法对跟踪目标视频前景图像序列提取具有互补性的视觉单词树直方图和全局颜色直方图二种特征,再采用多核学习方法在线训练人体目标视觉外观,从而得到多核特征融合模型.实验结果表明,该方法能快速训练人体目标外观模型,满足视频监控的实时要求,多核融合模型获得了比单一特征模型和单核支持向量机方法更高的识别性能.     

An Adaptive Superpixel Tracker Using Multiple Features

Visual tracking is a challenging issue in the field of computer vision due to the objects’ intricate appearance variation. To adapt the change of the appearance, multiple channel features which could provide more information are used. However, the low level feature could not represent the structure of the object. In this paper, a superpixel-based adaptive tracking algorithm by using color histogram and haar-like feature is proposed, whose feature is classified into the middle level. Based on the superpixel representation of video frames, the haar-like feature is extracted at the superpixel level as the local feature, and the color histogram feature is applied with the combination of background subtraction method as the frame feature. Then, local features are clustered and weighted according to the target label and the location center. Superpixel-based appearance model is measured by using the sum of the voting map, and the candidate with the highest score is selected as the tracking result. Finally, an efficient template updating scheme is introduced to obtain the robust results and improve the computational efficiency. The proposed algorithm is evaluated on eight challenging video sequences and experimental results demonstrate that the proposed method can get better performance on occlusion, illumination variation and transformation.     

基于卷积网络的目标跟踪应用研究

本文针对目标跟踪应用,提出了基于Siamese-FC跟踪网络的改进卷积网络Siamese-MF,意在更进一步提升跟踪速度和准确性,满足目标跟踪的工程应用需求。对于跟踪网络,考虑速度和精度的权衡,减少计算量,增加卷积特征的感受野是改进跟踪网络的速度和精度的方向。在卷积网络结构上面进行改进结构创新,改进主要集中为两点:1)引入特征融合,丰富特征;2)引入空洞卷积,减少计算量的同时增强感受野。Siamese-MF算法实现了对于复杂场景目标的实时准确跟踪,在公开数据集OTB上测试速度达到平均76 f/s,跟踪成功率的均值达到0.44,而跟踪稳定性的均值达到0.61,实时性、准确性和稳定性均提升,满足目标实时跟踪应用。     

基于声呐图像序列SIFT特征的多目标跟踪方法研究

针对高频成像声呐多目标跟踪中声成像的不稳定问题,在基于检测前跟踪的声呐图像序列尺度不变特征变换（Scale Invariant Feature Transform,SIFT）的基础上,提出了一种结合对数变换的声呐图像序列SIFT特征跟踪方法。该方法利用对数变换进行图像预处理,进而利用声呐图像序列SIFT特征进行多目标跟踪。水池试验数据的处理结果表明：与中值滤波和动态亮度分配的方法相比,结合对数变换的声呐图像序列SIFT特征跟踪方法具有更好的多目标跟踪效果。     

基于R-FCN框架的多候选关联在线多目标跟踪

在线多目标跟踪是实时视频序列分析的重要前提。针对在线多目标跟踪中目标检测可靠性低、跟踪丢失较多、轨迹不平滑等问题,提出了基于R-FCN网络框架的多候选关联的在线多目标跟踪模型。首先,通过基于R-FCN网络从KF预测结果和检测结果中获取更可靠的候选框,然后利用Siamese网络进行基于外观特征的相似性度量,实现候选与轨迹之间的数据关联,最后通过RANSAC算法优化跟踪轨迹。在人流密集和目标被部分遮挡的复杂场景中,提出的算法具有较高的目标识别和跟踪能力,大幅减少漏检和误检现象,跟踪轨迹更加连续平滑。实验结果表明,在同等条件下,与当前已有的方法对比,本文提出在目标跟踪准确度(MOTA)、丢失轨迹数(ML)和误报次数(FN)等多个性能指标均有较大提升。     

基于自适应特征融合的多尺度相关滤波跟踪

为了减小目标跟踪中目标变形、光照影响、运动模糊以及目标旋转对跟踪效果的影响,在相关滤波KCF基础上,提出了一种基于自适应特征融合的多尺度相关滤波跟踪算法。首先,提取VGG19网络中conv2-2、conv3-4、conv5-4层的特征以及CN特征,并在conv2-2层加入CN特征;然后,将这3个特征分别代替HOG特征进行滤波学习,得到3幅响应图;进而对3幅响应图进行加权融合预测目标位置。最后,在尺度方面引入多尺度相关滤波器进行尺度的确定。该算法比KCF跟踪算法精确度和成功率分别提高了13.6%和11.8%。与现有的其他优异跟踪算法相比,该算法在应对运动模糊、背景杂乱、目标变形、平面旋转方面更具有较好的跟踪效果。     

Electronic speckle pattern interferometry based on spatial fringe analysis method using two cameras

A new speckle measurement method is proposed by applying a spatial phase shifting method to multi-camera technology in order to perform a high resolution, high speed, and large deformation measurement. It is confirmed that the alignment of optical elements in this method is easier than the ordinary multi-camera methods because the optical system uses only two cameras. The validity of principle of the method is discussed by the results of experiments. It is shown that measurement precision of this method is about 1/50 wavelength in a small deformation measurement. Furthermore, the method is improved for a large deformation measurement method by accumulating the results of the small continuous deformation measurement. The optimum sampling process of the large deformation of an object is proposed in order to detect the phase map of the large deformation. It is confirmed that the large deformation can be precisely measured by this method.     

一种基于动态分层更新的多特征融合跟踪方法

针对单一特征的目标跟踪算法和传统的模型更新策略的不足,本文提出一种均值迁移和粒子滤波相结合的多特征融合跟踪方法.该方法通过均值迁移对粒子传播进行优化,根据粒子权值的分布情况自动调节各个特征的融合权值,实现了多特征的有效综合,通过建立目标模型的动态分层更新策略,有效保留了目标和场景的变化信息.实验结果表明,该方法对目标外...     

基于贝叶斯推理时空交互式多视角目标跟踪

本文针对多摄像机视频跟踪的应用场景,在贝叶斯推理的框架下,提出了一种具有分布式时空交互计算特点的目标跟踪算法.本文首先利用贝叶斯网络对拓扑确定的多摄像机系统进行建模,并对待估状态(目标位置)高阶联合后验概率密度函数进行时空的递推,最后借助序列蒙特卡洛(粒子滤波)逼近后验概率密度函数,并采用高效的数据传送机制高效求解出跟...     

Developing an RFID-based tracking system to improve the control of construction surplus soil disposal in Taiwan

The current tracking system for construction surplus disposal soil report system (CSRS) in Taiwan is labor-intensive and prone to human error, with its accuracy and effectiveness often questioned. Radio-frequency identification (RFID) is a wireless sensing technology that uses radio waves and signals to wirelessly transmit, retrieve, and store data to identify the status of objects and contents. It can be read at long ranges and be operated in extreme environments. This paper has improved the existing construction surplus soil tracking system by developing an integrated RFID-based tracking system (RFID-CSRS). To accommodate the characteristics of the domain problems, RFID-CSRS integrates RFID technology with Internet infrastructure, cameras, and signal controlling units. The proposed system reduces human errors, such as incorrect manual recording and identification, inaccurate monthly reports and audits, and improper picture-taking of trucks. RFID-CSRS was also assembled and tested in two surplus soil disposal cases in Taiwan. The results demonstrated the soundness of the system and its capability to improve the effectiveness of the current system in tracking the disposal of construction surplus soil.     

A domain-independent interaction integral for linear elastic fracture analysis of micropolar materials

This paper establishes a domain-independent interaction integral (DII-integral) for linear elastic fracture mechanics of micropolar elastic solids. The DII-integral has three amazing features that make it effective for solving the fracture parameters of complex micropolar materials. The first one is that the DII-integral can decouple the stress intensity factors (SIFs) and couple stress intensity factors (CSIFs) both of which are the key fracture parameters charactering the crack-tip asymptotic singular fields. In details, the DII-integral is derived from the J-integral by superimposing an actual field and an auxiliary field. By assigning the fracture parameters in the auxiliary field with different values, the SIFs and CSIFs of different crack opening modes can be obtained separately through the DII-integral. The second important feature is that the DII-integral is domain-independent for material nonhomogeneity and discontinuity. Thanks to this feature, the DII-integral becomes extremely effective for the micropolar materials with arbitrary nonhomogeneous properties or complex interfaces. The third feature is that the DII-integral does not contain any derivatives of material properties, which feature facilitate the practical implementation of the DII-integral on complex micropolar materials. Finally, the DII-integral combined with the extended finite element method (XFEM) is employed to solve four representative crack problems and the results show good validity of the DII-integral for complex micropolar materials.     

An Efficient Crossing-Line Crowd Counting Algorithm with Two-Stage Detection

Crowd counting is a challenging task in crowded scenes due to heavy occlusions, appearance variations and perspective distortions. Current crowd counting methods typically operate on an image patch level with overlaps, then sum over the patches to get the final count. In this paper we describe a real-time pedestrian counting framework based on a two-stage human detection algorithm. Existing works with overhead cameras is mainly based on visual tracking, and their robustness is rather limited. On the other hand, some works, which focus on improving the performances, are too complicated to be realistic. By adopting a line sampling process, a temporal slice image can be obtained for pedestrian counting without the need for visual tracking. Only ten low level features are extracted from the input image to establish a feature vector. As a result, our algorithm is more efficient and accurate than existing methods. Pedestrians in the temporal slice image are then located by the two-stage detection algorithm, which is largely based on support vector machine and affinity propagation clustering. Moreover, a novel algorithm is proposed to determine the moving directions of pedestrians by comparing the centers of them in two temporal slice images. Extensive experiments reveal that our system achieves satisfaction performances in terms of both robustness and efficiency.     

基于连续谱特征的被动声呐跟踪方法

针对被动声呐多目标跟踪问题,通过研究目标连续谱特征表征方式和特征更新相似关联机制,提出了一种基于连续谱特征的被动目标跟踪方法。该方法利综合利用频带能量法、排序截断平均算法以及峰值提取实现目标的连续特征表征,利用表征出的特征谱作为输入,通过建立相似度搜索,交叉判断与模板更新机制,实现了多目标的跟踪。经仿真对比分析了该方法的跟踪性能,并利用海上试验数据验证了其有效性。结果表明,连续谱特征可作为辅助特征用于目标的跟踪分辨,该方法能够有效提高多目标交叉情况下的跟踪关联能力,并且具备较低的运算量和较高的精度。     

基于局部特征组合的目标跟踪算法

为了克服目前大多数观测模型在小样本空间中鲁棒性不高的弱点,文中在粒子滤波框架下提出基于局部特征组合的粒子滤波视频跟踪算法。局部特征能更有效描述目标模板细节信息,可降低特征匹配中目标形变、光照变化和部分遮挡的影响。该方法借鉴混合高斯模型思想,采用多模式描述有效局部观测信息,这种融合策略更加准确可靠,能够较好地通过最新观测减轻了粒子退化现象,从而提高目标跟踪效率。小样本空间一定程度上降低了粒子数量和计算代价。实验结果表明该算法相比单一特征或一般多特征融合跟踪算法具有优越性,并能实现复杂场景下的目标跟踪。     

Three-dimensional motion measurements using feature tracking

Feature tracking was developed to efficiently compute motion measurements from volumetric ultrasound images. Prior studies have demonstrated the motion magnitude accuracy and computation speed of feature tracking. However, the previous feature tracking implementations were limited by performance of their calculations in rectilinear coordinates. Also, the previous feature tracking approaches did not fully explore the three dimensional (3- D) nature of volumetric image analysis or utilize the 3-D directional information from the tracking calculations. This study presents an improved feature tracking method which achieves further computation speed gains by performing all calculations in the native spherical coordinates of the 3-D ultrasound image. The novel method utilizes a statistical analysis of tracked directions of motion to achieve better rejection of false tracking matches. Results from in vitro tracking of a speckle target show that the new feature tracking method is significantly faster than correlation search and can accurately determine target motion magnitude and 3-D direction.     

Motion tracking system in video based on extensive feature set

In recent years, object detection and tracking has been a dynamic research area. Rapid development of the multimedia and the associated technologies urge the processing of a huge database of video clips. The processing efficiency lies on the search methodologies utilised in the video processing system. Usage of unsuitable search methodologies may make the processing system ineffective. Hence, effective object detection and tracking system is an essential criterion for searching relevant videos from a huge collection of videos. This paper proposes a unique object detection and tracking system where video segmentation, feature extraction, object detection and tracking are combined perfectly using various features. Initially, the database video clips are segmented into different shots before performing the feature extraction process. The proposed system consists of two stages, namely, feature extraction and tracking of object in the video clips. In the feature extraction stage, firstly, colour feature is extracted based on colour quantisation. Next, edge density feature is extracted for the objects present in the query video. Then, the texture feature is extracted based on LGXP technique. Finally, based on these feature extracted, the object will be detected and the detected objects will be tracked by utilising both forward and backward tracking technique. The proposed methodology proved to be more effective and accurate in object detection and tracking.     

Efficient object detection and tracking in video sequences

One of the most important problems in computer vision is the computation of the two-dimensional projective transformation (homography) that maps features of planar objects in different images and videos. This computation is required by many applications such as image mosaicking, image registration, and augmented reality. The real-time performance imposes constraints on the methods used. In this paper, we address the real-time detection and tracking of planar objects in a video sequence where the object of interest is given by a reference image template. Most existing approaches for homography estimation are based on two steps: feature extraction (first step) followed by a combinatorial optimization method (second step) to match features between the reference template and the scene frame. This paper has two main contributions. First, we detect both planar and nonplanar objects via efficient object feature classification in the input images, which is applied prior to performing the matching step. Second, for the tracking part (planar objects), we propose a fast method for the computation of the homography that is based on the transferred object features and their associated local raw brightness. The advantage of the proposed schemes is a fast matching as well as fast and robust object registration that is given by either a homography or three-dimensional pose.     

复杂背景及遮挡条件下的运动目标跟踪

CamShift算法应用于复杂背景及遮挡条件下视频跟踪时,极易出现跟踪失效和目标丢失。本文提出基于颜色、纹理及目标运动信息的综合特征用于改进CamShift算法,结合Kalman滤波器对目标运动状态进行预测提高了复杂背景下运动目标的跟踪稳定性和跟踪精度。在目标发生遮挡时,通过目标遮挡前的先验信息进行最小二乘拟合及目标运动轨迹外推,预测目标运动位置信息,有利于遮挡结束时对运动目标的重新捕获。多组实验结果及性能分析表明,该算法在复杂背景及目标被短时遮挡情况下,可以实现目标的持续、稳定跟踪,并具有较好的实时性。     

自适应多滤波器的高效卷积算子目标跟踪算法

针对单一滤波器难以适应复杂变化的目标跟踪环境的问题,本文在高效卷积算子目标跟踪算法的基础上,提出了自适应多滤波器的目标跟踪算法。该算法使用时空正则化滤波器、一致性检验滤波器和高效卷积算子算法中的相关滤波器分别与目标特征进行卷积,得到三个滤波检测得分。其中,时空正则化滤波器是通过将时间正则化引入相关滤波损失函数而得到;一致性检验滤波器是通过反向定位前几帧目标,比较反向与正向定位坐标的误差,只有误差小于阈值时才更新滤波器;选择峰值旁瓣比最大滤波检测得分,估计目标的位置。使用OTB-2015数据集和UAV123数据集对改进算法进行测试,实验结果表明,本文算法能够更好地适应跟踪过程中的复杂变化的环境,具有较高的精度和鲁棒性。