机械手视觉系统设计与实现

Design and realization of robot-vision system

本文介绍了一种基于智能摄像机的智能机器人视觉系统。它采用Blob分析法可以求取任意形状目标物的中心坐标及其任一斜边与坐标轴之间的夹角．从而实现对微小或缓慢运动的目标物定位。系统硬件平台是内含TMS320C64XX系列DSP的智能摄像机。利用VC智能摄像机来完成场景图像的采集及量化任务,并利用其自带的嵌入式操作系统及开发平台编写了相应的目标定位程序。实验结果表明,该系统能满足实际项目中对目标姿态信息的需要,目标识别率达到100％,定位相对误差小于5％。本系统具有广泛的应用前景．通过修改应用程序可将该系统应用于不同的工业机器视觉检测领域．尤其是图像测量、表面检测．汽车制造业检测、产品包装和印刷业的质量检测、运动轨迹分析与跟踪等领域。

A robot-vision system applied for detecting small and slow moving objects is introduced. It utilizes Blob-analysis method to recognize the center of objects with arbitrary form and the angle between axis and arbitrary hypotenuse of objects. An intelligent camera with TMS320C64X series DSP is adopted as the hardware platform of the system. The VC intelligent vidicon is used to fulfill the collection and quantification of scene picture data, and an embedded -operating system and developing platform are involved to implement the corresponding target-localization program. The simulating experiments indicated that the system can meet the requirements of system design, the target recognition rate can reach 100%, and the localization relative error is less than 10%. The system can be widely used to the different machine vision fields in industry with different application programs, especially in picture detection, superficial examination, automotive industry examination and so on.