A Self-Calibrated Pipeline ADC with 200 MHz IF-Sampling Frontend

A 13-bit, 50-MS/s pipeline ADC with IF-sampling capability is presented. A high sampling linearity is obtained through the use of bootstrapped switches. A digital self-calibration algorithm with modified capacitor measurement scheme is employed to improve the accuracy of the first two pipeline stages. The prototype, implemented with a 0.35-m BiCMOS (SiGe) technology, shows a 76.5-dB SFDR at a 194.2-MHz signal frequency and dissipates 715 mW power from a 2.9-V supply.     

基于立体视觉的人体头部目标识别研究

使用立体视觉系统可确定任意物体的三维轮廓及其表面任意点的位置信息和深度信息。本文将立体视觉运用于人体头部目标检测,设计了一种基于图像深度信息处理的人体头部识别系统。系统采用Xtion摄像机采集场景的深度图像,并对其进行特征分析,根据深度图像的特点以及头部的特征确定头部目标区域。再对目标区域采用Mean Shift算法进行聚类处理,得到清晰的图像边缘。最后通过基于动态阈值的一维熵函数分割法实现头部的分割识别。该系统可快速锁定目标区域,减少了算法的计算量,大大提高了系统的识别速度。此外,采集深度图像的xtion摄像机悬挂在目标场景的正上方,因而较好的解决了目前人体目标检测受遮挡的问题。经实验论证,该系统有较高的识别精度。     

ADV-JP2000及其在数码相机中的应用

ADV -JP2000是AnalogDevice公司为新的静止图像压缩标准JPEG2000而生产的专用压缩芯片 ,该芯片可灵活应用于各种静止图像压缩场合。文中主要介绍了ADV -JP2000的工作过程 ,给出了用ADV -JP2000为处理核心的数码相机系统的设计方案     

基于双目立体视觉的坝体形变测量方法

GPS坝位移动自动化检测系统是利用现代化电子信息技术、通信技术及计算机技术,实现对坝体位移检测指标数据实时、自动、连续采集、传输、管理、分析及预警的监测技术。本文设计了一种基于双目立体视觉的坝体形变测量方法。该方法采用双目立体视觉的测量原理,结合图像预处理、图像分割、角点检测、立体匹配的方法,计算设置在坝体上特征点的三维空间坐标,从而获取坝体形变信息。仿真实验表明,该坝体形变测量方法能实时显示坝体形变数据,并且其可视化的输出结果能够直观显示坝体形变状态。     

摄像机标定方法综述

首先介绍了摄像机标定的基本原理以及对摄像机标定方法的分类。通过对最优化标定法、双平面标定法、两步法等传统摄像机方法的具体分析，给出了各种方法的优劣对比；同时对多种自标定方法的研究现状、发展情况以及存在问题进行了探讨。最后给出了发展传统摄像机标定方向、提高摄像机自标定精度的二些参考建议。     

基于变分贝叶斯估计的相机抖动模糊图像的盲复原算法

在曝光过程中由于相机抖动而导致的运动模糊,是一种常见的图像降质现象。该文提出了一种基于变分贝叶斯估计和自然图像梯度统计特性的盲复原算法,用于恢复相机抖动模糊图像,同时针对图像复原过程中出现的振铃效应,设计了一种基于分区域检测和Fuzzy滤波器的去振铃效应方法。实验结果表明,该文提出的盲复原算法能够有效地去除图像中因相机抖动而产生的模糊,而且在保持图像边缘和细节的同时,可以较好地降低振铃效应对图像复原质量的影响。     

无线传感器网络RSSI自校正定位

无线传感器网络的许多应用都和节点的位置紧密相关,节点自定位成为当前研究的一个热点。针对信号强度受环境影响较大,提出了接收信号强度自校正定位算法。该算法利用信标节点实际测量信号确定无线信号传播损耗模型,采用组合3边测量获得节点坐标估计值集合,再用加权质心方法确定节点的最终坐标。仿真结果表明,该算法较传统的基于接收信号强度定位方法定位精度有了明显的改进。     

数字视频系统接口应用

随着技术发展,数字视频接口逐步替代模拟视频接口,成为视频产生、传输、显示、存储的主要方式.列举了常用的模拟和数字接口,阐述了视频系统前端与后端数字视频接口的特点及应用情况.通过分析各个端口的性能和实际军事需要,得出Camera Link接口是现在军用光电产品中的主要发展和应用方向.并且从接口的发展趋势来看,未来的数字接口是向摄取速度更高、传输带宽更宽、分辨率更高的方向发展.     

ITER CODAC interface for the visible and infra-red wide angle viewing cameras

The amount of data generated by the infra-red and visible cameras at ITER is expected to be considerably larger than most diagnostics. ITER will have 12 infra-red cameras plus 12 visible cameras in four different equatorial port plugs. Each of the ports will have a Plant System Host (PSH) that will provide a standard image of the plant system to the ITER's Control and Data Access and Communication (CODAC) system.The two key functions of these cameras will be the scientific exploitation with the detection of interesting physics events and the operational protection of the machine, namely the first wall. Already assuming high bandwidth requirements for both audio and video, ITER will provide a separate network for this kind of communication, which will be used to transmit both the experimental and informational data provided by the cameras.This paper presents the current camera plant system design and its interaction with ITER CODAC system and networks. Starting from the camera specifications several alternatives for data collection and compression are discussed. The required inputs from CODAC and a first specification for the internal finite state machine are also presented. Finally, a possible hardware straw man design solution for the plant system hardware is proposed.     

Enhancing pose accuracy of space robot by improved differential evolution

Due to the intense vibration during launching and rigorous orbital temperature environment, the kinematic parameters of space robot may be largely deviated from their nominal parameters. The disparity will cause the real pose (including position and orientation) of the end effector not to match the desired one, and further hinder the space robot from performing the scheduled mission. To improve pose accuracy of space robot, a new self-calibration method using the distance measurement provided by a laser-ranger fixed on the end-effector is proposed. A distance-measurement model of the space robot is built according to the distance from the starting point of the laser beam to the intersection point at the declining plane. Based on the model, the cost function about the pose error is derived. The kinematic calibration is transferred to a non-linear system optimization problem, which is solved by the improved differential evolution (DE) algorithm. A six-degree of freedom (6-DOF) robot is used as a practical simulation example, and the simulation results show: 1) A significant improvement of pose accuracy of space robot can be obtained by distance measurement only; 2) Search efficiency is increased by improved DE; 3) More calibration configurations may make calibration results better.