基于压缩域互补分割的视频多描述编码研究

鉴于数字视频在空间域及时间域上的信息相关性,提出在视频压缩过程中实现多描述编码的方法.算法以H.263为参考标准,在图块及运动估计的DCT域上进行系数划分,通过对少量重要信号的重复分配来保证单个描述的接收质量.同时,系数域的对角分割又使各个描述间具有较强的信号互补性.实验结果表明,算法具有描述冗余率低、编/解码直接快速、压缩率高等良好的MDC特性,是保障视频传输业务可靠性和实时性的有效方法.     

生物Reichardt运动检测器和感受域模板的 FPGA设计与实现

为了在图像处理中快速地实现运动检测和相机自身运动方式估算,引入了基于生物视觉机制的Reichardt运动检测器模型(EMD)和感受域模板.分析了Reichardt运动检测器模型的基本特性及其缺陷.为了克服模型上的主要缺陷,在应用中选择了一种优化模型,应用该模型可以得到较好的运动检测结果.同时,提出了基于苍蝇视觉系统的6个感受域模板,用以实现简单自身运动方式的估算,如相机自身的平移、旋转等.最后,在FPGA(FieldProgrammable Gate Array)平台上实现了相关的算法.实验结果表明,优化后的运动检测器可以快速地判断局部运动方向,感受域模板可实现在特定背景下的简单运动方式估算;对分辨率为256×256像素的输入图片,本设计中的FPGA系统可达到每秒350帧的处理速率,所产生的延时仅为0.25μs,达到了快速处理的要求.此模型可应用于实时的机器视觉系统,如进行障碍物检测、整体运动方式估算、UAV/MAV的稳定控制等.     

一级直线倒立摆匀速行走的模糊控制研究与实现

当前针对倒立摆的研究一般是把角度控制或者位置控制作为控制目标,很少着眼于速度控制,有鉴于此,设计了一种简单的模糊控制器,应用于一级直线倒立摆的匀速行走控制中;仿真实验和实物实时控制实验均验证了该控制器的有效性,对于设定的速度整定值,通过调整几个比例环节的系数,系统具有很好的动态性能指标,而且控制器在有外界扰动时体现了很好的抗扰性;考虑到实际物理系统中倒立摆行程的限制,设计的自动换向开关实现了倒立摆在一段给定的行程上匀速来回行走的控制目标,通过手动切换开关也能实现倒立摆的位置控制。     

LCD motion‐blur estimation using different measurement methods

Abstract— The primary goal of this study was to find a measurement method for motion blur which is easy to carry out and gives results that can be reproduced from one lab to another. This method should be able to also take into account methods for reduction of motion blur such as backlight flashing. Two methods have been compared. The first method uses a high‐speed camera that permits us to directly picture the blurred‐edge profile. The second one exploits the mathematical analysis of the motion‐blur formation to construct the blurred‐edge profile from the temporal step response. Measurement results and method proposals are given and discussed.     

Moving‐picture response time (MPRT) of LCDs for the oblique viewing direction

Abstract— The dynamic performance of displays is an important characteristic for multimedia applications. Motion‐picture response time (MPRT) has been used as an indicator of the dynamic performance of LCDs. This paper describes a comprehensive method of MPRT evaluation for the oblique viewing direction. By using a tilted camera configuration, the angular dependency of MPRT is investigated for the condition that the horizontally scrolling patterns are observed from the vertical direction. For each gray‐to‐gray transition, distinct changes in MPRT and the luminance profile of blur are observed.     

Biomechanical evaluation of bike power saver

Bike power saver (BPS) is a new product that claims to change the angle of pedaling forces and reduce applied power dead range (Chic Sheng Industrial Co., Ltd, Taiwan, China). In order to determine its effectiveness, we quantified how BPS operates through a 3D kinematical study and electromyography (EMG) analysis of leg muscles during pedaling. Ten kinesiology students participated in this study. A 3D motion capture system consisting of nine high-speed cameras (VICON v8i, a capture rate of 120 frames/s) was used to collect the total body and pedal motion with and without BPS at statically determined low, middle and high intensity cycling levels. The short-time test (14s) was applied to all intensity levels while the long-time test (30 min) was applied only to the low wattage level. Wireless EMG was synchronized with the 3D motion capture system to monitor the right and left tibialis anterior, gastrocnemius, quadriceps and hamstring. The results revealed that BPS did not alter hip and knee movement significantly (p>0.05), but it did vary ankle movement. BPS caused a movement change in the pedals, and consequently induced instability in ankle control. The altered pedal movement led to an increase in activity level and presumably also energy expenditure for dominant muscles, resulting in a faster fatigue process. From these data, it is likely that the BPS actually requires more effort than a standard bike.     

排球机器人开放式控制系统的设计

为了提高中国排球运动员的训练水平,提出一种用于排球扣球机器人的开放式控制系统。通过该控制系统,排球扣球机器人可以有效地辅助教练员和运动员进行日常训练,提高运动员的拦网和一传训练水平。针对传统机器人采用封闭式控制系统的局限性,将开放式控制系统的思想引入到排球机器人控制系统的研制中。实际运行结果表明,该排球扣球机器人具有运行稳定、控制准确和灵活的特点。     

基于子空间集成学习的3维人体运动识别

如何对3维运动数据进行自动识别,是提高数据利用效率和进行计算机动画创作的前提。为了有效地识别运动数据,需要把运动数据投影到非线性流型低维子空间中,先识别出运动的内在结构,然后对运动的各个关节点分别进行学习,最后基于集成学习的方法产生强的隐马尔可夫学习器,以便能够对一些常见的运动类型进行自动识别。实验结果表明,这种基于子空间集成的人体运动识别方法较传统方法的检索精度、检索速度均有较大提高。     

Neural reactive path planning with Riemannian motion policies for robotic silicone sealing

Due to its excellent chemical and mechanical properties, silicone sealing has been widely used in many industries. Currently, the majority of these sealing tasks are performed by human workers. Hence, they are susceptible to labor shortage problems. The use of vision-guided robotic systems is a feasible alternative to automate these types of repetitive and tedious manipulation tasks. In this paper, we present the development of a new method to automate silicone sealing with robotic manipulators. To this end, we propose a novel neural path planning framework that leverages fractional-order differentiation for robust seam detection with vision and a Riemannian motion policy for effectively learning the manipulation of a sealing gun. Optimal control commands can be computed analytically by designing a deep neural network that predicts the acceleration and associated Riemannian metric of the sealing gun from feedback signals. The performance of our new methodology is experimentally validated with a robotic platform conducting multiple silicone sealing tasks in unstructured situations. The reported results demonstrate that compared with directly predicting the control commands, our neural path planner achieves a more generalizable performance on unseen workpieces and is more robust to human/environment disturbances.