具有3-2下拉检测功能的运动自适应去隔行算法

提出了一种从隔行扫描到逐行扫描的格式转换算法．该算法首先计算相隔一场的两个电视场亮度信号之间的差值，即相邻的两个奇场(或偶场)亮度信号之间的隔场差。然后，对差值信号执行形态开运算，以滤除一些高频噪声。通过使用隔场差，不仅能正确检测静止区域和运动区域，而且还能检测经3—2下拉变换后的电影序列.对于检测到的电影序列，则进行3—2下拉恢复处理，输出电影模式的逐行扫描帧；否则，采用运动自适应的插补方法进行去隔行处理，转换成具有高垂直分辨率的逐行扫描帧图像。     

结合前景轮廓提取的改进ViBe运动目标检测算法

运动目标检测是对视频或图像序列进行一系列地分析运算,计算出运动目标在图像中的位置,并将运动目标从复杂背景中提取出来,得到只包含运动目标的图像. ViBe是目前常用的运动目标检测算法,但使用ViBe算法对视频或图像序列中的运动目标进行识别与检测时,存在“鬼影”现象并易受环境噪声的影响,为此,设计了前景轮廓提取算法对ViBe算法改进.构建背景模型时,前景轮廓提取算法使用众数背景建模法建立初始背景及背景序列;前景检测时,使用背景差分法和Sobel算子计算出运动目标区域,使用自适应的多级阈值去噪;最后,将前景轮廓提取算法与ViBe算法求交集,并使用数学形态学处理获取完整的运动目标.同时,设计了CPU-GPU的并行方案,使用CPU并行的计算图像背景,使用GPU加速前景检测.将算法在CDNet2014数据集上进行测试,实验结果表明,算法的检测精确率、召回率、F1分数较ViBe分别提高了32.14%、9.64%、20.76%,漏检率及错检率较低,精度较高;效率方面,算法的平均检测帧率较ViBe算法提升了64.70%,具有较好的实时性.     

利用相邻帧和背景信息的运动对象检测

 运动对象检测是智能视频监控的基础,现有的运动对象检测算法在实时性和检测精度方面仍有待提高,本文在分析相邻帧间的亮度变化基础上,提出了基于三阶中心矩统计分布的背景点阈值计算方法,给出了抗噪声干扰的背景维护模型,最后综合利用时间差分法和背景减法进行运动对象检测,为了提高检测精度使用了形态学的腐蚀与扩张处理.对室内、室外序列以及标准序列的测试结果表明,本文算法速度快、抗噪声能力强,能够检测出被遮挡的运动对象.     

一种微弱点运动目标的快速统计检测算法

本文提出一种新的从序列图像中检测微弱点状运动目标算法,该算法基于多帧检测技术,它首先累加多个差分帧,构成组合帧,然后在组合帧内对目标轨迹进行统计判别.与其它检测算法相比较,该算法在保证较高检测性能的同时,具有快速、实时的特点.本文还给出了该算法性能的理论分析结果及实验仿真结果.     

基于3D OGHM的视频运动目标检测算法

本文提出了一种新的视频运动目标检测算法.该算法联合了视频序列图像特有的时、空特性,采用3D Gassian-Hermite正交矩法(3D OGHM,3D Orthogonal Gassian-Hermite Moments)进行视频序列中的运动目标检测.并对提出的算法进行了深入的分析和优化,给出了3D OGHM算法的具体实现步骤.实验结果表明:算法能够从复杂场景的视频序列中有效地检测出运动目标;与同类算法相比,不仅有更好的抗噪性能,而且能更有效地增强运动目标信息,且具有较强的鲁棒性和实时性.     

基于图像序列的VideoSAR动目标检测方法

VideoSAR成像时,运动目标发生散焦和移位,但在其真实位置留下阴影,且阴影在VideoSAR的图像序列中发生运动。根据运动目标的这个特征,提出了一种在VideoSAR图像序列中检测运动目标的方法,通过检测运动的阴影来检测运动目标。该方法主要分为背景补偿和背景分割两部分。首先,将段内的图像用SIFT+RANSAC算法进行配准,使得图像序列的背景对齐;然后,用单高斯模型对图像序列进行统计,得到该段图像序列的背景模型,将背景与当前图像差分后并二值化,得到前景的二值图;最后,对二值图进行形态学处理,即可提取出运动目标的阴影。该方法用美国Sandia国家实验室的VideoSAR图像进行了验证,具有良...     

基于区域分割的噪声抑制及红外目标检测

提出一种新的基于区域分割与邻域梯度差平方累积最小(MCSE)原则的自适应滤波算法.该算法对各种加性、乘性噪声的滤波性能优异且均衡,适合滤除图像序列中的复杂噪声.滤波后,采用改进光流方程算法(OFE)估计图像序列的运动场,为避免传统算法在图像补偿时引入的误差,本文直接对运动场进行全域运动补偿,然后在运动场中引入一种新的集成检测算法对微小目标的运动矢量进行检测.理论分析及试验表明,在复杂噪声干扰下,本算法依然能够在具有全域运动的图像序列中检测微小目标.     

高速密集视频目标场景下的运动分析

本文首先分析了高速、密集图像目标场景下邻帧差分运动检估计检测法的缺陷.然后综合图像能量、高阶统计量预测理论以及块处理技术,对高速密集图像序列进行快速背景重建、运动区域提取与运动分析.基于DSP的平台测试表明:本文背景重建算法速度快,运动区域检测精度高,目标跟踪准确.为密集、高速的图像目标跟踪系统的研制提供了新的可借鉴方法.     

Novel De-interlacing Algorithm Based on Motion-adaptive Detection

针对隔行视频,提出了一种新的运动自适应去隔行算法,该算法能充分利用图像序列时间和空间上的相关信息,采用新颖的运动检测方法区分视频序列中的运动和静止部分.对于运动部分,使用新的自适应插值算法;对于静止部分,采用行复制算法进行处理.实验表明,该算法与一些传统的去隔行算法相比,在画面视觉效果上,能达到更好的效果.     

基于水平集的热红外运动人体目标分割算法

水平集活动轮廓模型是一种优秀的图像分割方法.针对红外人体检测系统中的图像分割难题,提出了一种基于水平集活动轮廓模型的新算法.该算法包含水平集运动检测模块、水平集亮度检测模块和融合模块.水平集运动检测模块融合了水平集和背景相减技术,通过演化水平集函数同时实现前景分割和背景估计,它用于检测序列中的运动区域,并将其演化结果输入到下一检测模块.水平集亮度检测模块融合了水平集和阈值分割技术.在给出双阈值时,可分割出亮度在双阈值所限定范围内图像区域,它用于检测序列图像序列中可能包含人体目标的全部区域.利用形态学开重建技术,融合模块在融合前两个模块检测结果后输出算法最终分割结果.此外,采用快速数值算法演化水平集检测模块以及优化设置整个算法流程,改善算法运行效率.实验结果表明,相对其他典型算法,该算法具有较高分割精度和运行效率,且对时序亮度变化和镜头运动鲁棒性更好.     

Algorithm analysis and architecture design for rate distortion optimized mode decision in high definition AVS video encoder

There are abundant intra and inter prediction modes in the AVS video coding standard. Rate distortion optimized mode decision can fully utilize this flexibility to improve the spatio-temporal prediction efficiency and maximize the coding efficiency. However, the implementation complexity is dramatically high due to huge throughput burden. Hardware oriented mode decision algorithm is tailored for VLSI implementation in this work for high definition video coding. Mode preselection is employed to alleviate the dramatic throughout burden. Also, intelligent pipeline scheduling mechanism is proposed to break the intrinsic data dependency in intra prediction, which is directly related with mode decision. The proposed simplified algorithm is well-suited for hardware implementation with small performance penalty. Finally, the VLSI architecture is proposed with good trade off between circuit consumption and rate distortion performance.     

一种H.264帧内预测模式判决算法及VLSI实现体系

17种预测模式和率失真优化模式判决极大的增加了H.264帧内编码器硬件设计的复杂度.目前的模式判决快速算法能大量减少模式判决的复杂度,但却不易于硬件实现.本文在Sobel边缘检测模式判决算法的基础上,提出了一种面向VLSI实现的模式判决优化算法.该算法通过修改16×16宏块部分像素的Sobel边缘检测算子来减少存储器读取次数,优化预测模式区域的范围来减少硬件设计复杂度,并采用变换后残差绝对值和(SATD)来简化编码代价判决运算.实验结果表明,采用该算法的帧内硬件编码器可以在确保编码质量的前提下,显著降低硬件实现复杂度和提高编码器效率.     

复合高斯杂波下的3种目标检测算法研究

针对复合高斯分布背景下的慢速目标检测问题,从贝叶斯统计检测理论出发,系统分析了最优的GLRT检测算法、次优的基于AR模型的目标检测算法和传统的自适应MTI检测算法。结合这3种算法的特点阐述了它们的检测性能和工程可实现性。并在最后通过IPIX雷达实测海杂波数据对上述结论进行了验证,实验结果表明在性能损失不大时,基于AR模型的目标检测算法易于工程实现。     

Single mode grating coupling between thin-film and fiber optical waveguides

Optical directional coupling by means of a periodic perturbation is described between a clad, single mode, cylindrical fiber, and a thin-film planar waveguide. Codirectional coupling was achieved to a clad fiber, with 0.4 percent efficiency; allowing for the geometric mismatch, this is effectively 30 percent coupling. The fiber guided mode is accessed by reducing the fiber diameter by heating and pulling. Mode phase matching is achieved with a periodic grating sputter etched into the film. Unwanted grating induced interactions are minimized by restricting the guide parameters or by modifying the coupling mechanism. A perturbation analysis is used to calculate coupling coefficients and the design procedure for an optimum structure is explained. Measurements on grating couplers are described and the film/fiber codirectional coupling is shown to be effected by the grating.     

光照对ADS模式TFT-LCD Flicker漂移的影响

在挠曲电效应的基础上,研究了光照对ADS模式TFT-LCD Flicker漂移的影响。首先研究了正常点灯下Flicker漂移情况,然后测试了不加背光和降低背光亮度下的Flicker漂移,最后研究了光照panel一段时间后Flicker漂移情况。实验结果表明,光照使Cell盒内产生了离子,驱动信号电压存在时液晶发生极化,产生的"力"吸附离子,从而形成了直流偏压,产生了Flicker漂移。降低光照强度Flicker漂移程度减弱。利用该结果在一定程度上对显示器整体设计和品质的改善有重要指导意义。     

Analysis and design of clamped-mode resonant converters withvariable load

In this paper, a design procedure for clamped-mode resonant converters working with variable load is proposed and analyzed. The operation of these converters with general parallel or series loads is reviewed first. The operation of the transistors as zero-voltage switches is identified and characterized as mode A. Mode A is preferred to simplify the implementation of the power switches and to increment the converter efficiency. As a case study, the design of an RF power amplifier is presented. The amplifier is connected to a variable load with reactive characteristics, through a low-pass filter of three elements. The validity of the proposed design approach is verified with an experimental setup     

基于FPGA的S模式信号处理

S模式信号处理终端是询问/应答设备的数据处理核心。结合实际工程应用,提出了一 种基于FPGA的S模式信号处理实现方案。针对S模式询问/应答信号的调制方式和信号格式, 给出了一种中频信号处理实现方法和信号的同步检测以及信息提取方式,通过对数据信息编 码/译码的原理分析,介绍了信号处理流程设计和实现方法。该方案已经过大量的仿真测 试并通过了工程应用验证。     

Low complexity VLSI implementation of CORDIC-based exponent calculation for neural networks

This article presents a low hardware complexity for exponent calculations based on CORDIC. The proposed CORDIC algorithm is designed to overcome major drawbacks (scale-factor compensation, low range of convergence and optimal selection of micro-rotations) of the conventional CORDIC in hyperbolic mode of operation. The micro-rotations are identified using leading-one bit detection with uni-direction rotations to eliminate redundant iterations and improve throughput. The efficiency and performance of the processor are independent of the probability of rotation angles being known prior to implementation. The eight-staged pipelined architecture implementation requires an 8?×?N ROM in the pre-processing unit for storing the initial coordinate values; it no longer requires the ROM for storing the elementary angles. It provides an area-time efficient design for VLSI implementation for calculating exponents in activation functions and Gaussain Potential Functions (GPF) in neural networks. The proposed CORDIC processor requires 32.68% less adders and 72.23% less registers compared to that of the conventional design. The proposed design when implemented on Virtex 2P (2vp50ff1148-6) device, dissipates 55.58% less power and has 45.09% less total gate count and 16.91% less delay as compared to Xilinx CORDIC Core. The detailed algorithm design along with FPGA implementation and area and time complexities is presented.     

Field-programmable analogue arrays for the sensorless control of DC motors

This work presents the analogue implementation of a sensorless controller for direct current motors based on the super-twisting (ST) sliding mode technique, by means of field programmable analogue arrays (FPAA). The novelty of this work is twofold, first is the use of the ST algorithm in a sensorless scheme for DC motors, and the implementation method of this type of sliding mode controllers in FPAAs. The ST algorithm reduces the chattering problem produced with the deliberate use of the sign function in classical sliding mode approaches. On the other hand, the advantages of the implementation method over a digital one are that the controller is not digitally approximated, the controller gains are not fine tuned and the implementation does not require the use of analogue-to-digital and digital-to-analogue converter circuits. In addition to this, the FPAA is a reconfigurable, lower cost and power consumption technology. Simulation and experimentation results were registered, where a more accurate transient response and lower power consumption were obtained by the proposed implementation method when compared to a digital implementation. Also, a more accurate performance by the DC motor is obtained with proposed sensorless ST technique when compared with a classical sliding mode approach.     

CMA Performance in CDMA Transmit Diversity System

1IntroductionThe capacity and performance of CDMAare largelyaffected by disturbances ,and Multi-Access Interference( MAI) is the main disturbance factor whenthe numberof user is large . Blind adaptive Multi-User Detection(B-MUD)is aneffectivetechniquetoreduce MAI[1 ~2].In many advanced and convergence faster algorithms ,Constant Mode Algorithm ( CMA) is an efficient B-MUDand has better systemperformance .Refs .[3 ~6]depict the CMAin details .The channel of mobile communications p…