基于双重注意力孪生网络的实时视觉跟踪

为了解决全卷积孪生网络（SiamFC）跟踪算法在跟踪目标经历剧烈的外观变化时容易发生模型漂移从而导致跟踪失败的问题，提出了一种双重注意力机制孪生网络（DASiam）去调整网络模型并且不需要在线更新。首先，主干网络使用修改后表达能力更强的并适用于目标跟踪任务的VGG网络；然后，在网络的中间层加入一个新的双重注意力机制去动态地提取特征，这种机制由通道注意机制和空间注意机制组成，分别对特征图的通道维度和空间维度进行变换得到双重注意特征图；最后，通过融合两个注意机制的特征图进一步提升模型的表征能力。在三个具有挑战性的跟踪基准库即OTB2013、OTB100和2017年视觉目标跟踪库（VOT2017）实时挑战上进行实验，实验结果表明，以40 frame/s的速度运行时，所提算法在OTB2013和OTB100上的成功率指标比基准SiamFC分别高出3.5个百分点和3个百分点，并且在VOT2017实时挑战上面超过了2017年的冠军SiamFC，验证了所提出算法的有效性。     

基于卷积神经网络优化TLD运动手势跟踪算法

针对TLD（Tracking-Learning-Detection）算法在光照变化不均、遮挡严重、跟踪目标模糊等情况下会出现跟踪失败的问题，提出一种基于卷积神经网络优化TLD运动手势跟踪算法。选取手势特征作正样本，其背景作负样本，获取手势HOG特征并投入到卷积神经网络中加以训练，得到手势检测分类器，从而确定目标手势区域，实现手势的自动识别；再利用TLD算法对手势进行跟踪与学习，对正负样本进行估计检测并实时校正，同时运用SURF特征匹配更新跟踪器。实验结果验证，该算法对比TLD经典算法跟踪精度提高了4.24%，增强了运动手势的跟踪效果，相比经典跟踪算法拥有更高鲁棒性。     

基于龙格-库塔法自适应PID控制的太阳双轴跟踪系统

论述太阳光源双轴跟踪系统的工作原理,针对其控制系统中存在的时间滞后问题提出基于龙格-库塔法的自适应PID控制。并以基于直流电机的方位角传动机构为基础,建立方位角传动机构的数学模型,在MATLAB中进行仿真。仿真结果表明,基于龙格-库塔法的自适应PID控制系统响应速度快、稳定性高,具有较强的鲁棒性,能有效解决方位角传动系统中的时间滞后问题。     

基于目标跟踪的变电站隔离开关状态图像识别方法

为解决变电站中隔离开关状态的自动识别问题，提出了基于目标跟踪的开关分合状态图像识别方法。首先，基于隔离开关开合过程的动态视频，通过 Ｌａｓｓｏ正则化稀疏表示的目标跟踪模型对隔离开关运动轨迹进行识别；其次，利用隔离开关２个闸刀之间的距离自动识别算法获得开关的分合状态。实测结果表明，该方法可快速跟踪开关运行轨迹，实时对变电站中隔离开关状态进行准确识别。     

基于船载卫星通信三轴天线的跟踪搜索技术

针对船载卫星通信三轴天线跟踪系统,介绍了船载三轴天线的系统组成,提出了一种基于捷联惯导的船载三轴天线跟踪搜索技术。天线以目标地理角为中心,通过坐标转换将目标地理角转换为天线的目标甲板角,调整AEC三轴使其及时准确地跟随目标甲板角。目前,该方法已经成功应用于多种船载三轴天线。实践表明,该方法能够快速准确地搜索到目标卫星,并实现对目标卫星的稳定跟踪。     

Looking ahead: Joint small group detection and tracking in crowd scenes

Small group detection and tracking in crowd scenes are basis for high level crowd analysis tasks. However, it suffers from the ambiguities in generating proper groups and in handling dynamic changes of group configurations. In this paper, we propose a novel delay decision-making based method for addressing the above problems, motivated by the idea that these ambiguities can be solved using rich temporal context. Specifically, given individual detections, small group hypotheses are generated. Then candidate group hypotheses across consecutive frames and their potential associations are built in a tree. By seeking for the best non-conflicting subset from the hypothesis tree, small groups are determined and simultaneously their trajectories are got. So this framework is called joint detection and tracking. This joint framework reduces the ambiguities in small group decision and tracking by looking ahead for several frames. However, it results in the unmanageable solution space because the number of track hypotheses grows exponentially over time. To solve this problem, effective pruning strategies are developed, which can keep the solution space manageable and also improve the credibility of small groups. Experiments on public datasets demonstrate the effectiveness of our method. The method achieves the state-of-the-art performance even in noisy crowd scenes.     

Artificial bee colony based algorithm for maximum power point tracking (MPPT) for PV systems operating under partial shaded conditions

Artificial bee colony (ABC) algorithm has several characteristics that make it more attractive than other bio-inspired methods. Particularly, it is simple, it uses fewer control parameters and its convergence is independent of the initial conditions. In this paper, a novel artificial bee colony based maximum power point tracking algorithm (MPPT) is proposed. The developed algorithm, does not allow only overcoming the common drawback of the conventional MPPT methods, but it gives a simple and a robust MPPT scheme. A co-simulation methodology, combining Matlab/Simulink™ and Cadence/Pspice™, is used to verify the effectiveness of the proposed method and compare its performance, under dynamic weather conditions, with that of the Particle Swarm Optimization (PSO) based MPPT algorithm. Moreover, a laboratory setup has been realized and used to experimentally validate the proposed ABC-based MPPT algorithm. Simulation and experimental results have shown the satisfactory performance of the proposed approach.     

Sag in drying coatings: Prediction and real time measurement with particle tracking

Sag is a coating phenomenon characterized by gravity-driven flow after deposition; excessive amounts of sag can lead to coating defects. In this work, a new method for evaluating and quantifying sag is investigated. The motion of micron-sized Lycopodium spores on an inclined coating surface is tracked during drying, and the resulting surface velocity data is used to determine sag length. This in situ particle tracking method is minimally invasive and permits real time measurements. Measured sag lengths and real time surface velocities in aqueous polyvinyl alcohol solution coatings compare well with a theoretical model. The model is also used to develop a predictive sag regime map, which anticipates the extent of sag given coating properties and process-specific parameters. This map also identifies viable processing windows and aids in intelligent coating design given specific process constraints. The predictions of the sag regime map are compared against experimental sag results from polyvinyl alcohol solution coatings as well as four commercial latex paints, revealing good agreement for coatings with Newtonian or ‘Newtonian-like’ rheologies.     

Utilizing sensor data to model students’ creativity in a digital environment

While creativity is essential for developing students’ broad expertise in Science, Technology, Engineering, and Math (STEM) fields, many students struggle with various aspects of being creative. Digital technologies have the unique opportunity to support the creative process by (1) recognizing elements of students’ creativity, such as when creativity is lacking (modeling step), and (2) providing tailored scaffolding based on that information (intervention step). However, to date little work exists on either of these aspects. Here, we focus on the modeling step. Specifically, we explore the utility of various sensing devices, including an eye tracker, a skin conductance bracelet, and an EEG sensor, for modeling creativity during an educational activity, namely geometry proof generation. We found reliable differences in sensor features characterizing low vs. high creativity students. We then applied machine learning to build classifiers that achieved good accuracy in distinguishing these two student groups, providing evidence that sensor features are valuable for modeling creativity.