摘要
针对微动并联机器人工作空间小、位姿测量时安装传感器比较困难的情况,提出了一种基于双目视觉原理的检测方法。根据图像坐标和世界坐标的对应关系,从两幅对应图像中抽取必要的特征点的信息,建立了基于特征点匹配的三自由度微动平台的位姿测量模型。这种求解位姿的方法,信息量少,运算量小,处理速度快,所需时间短,尤其适于动态测量。测量系统由两台较高分辨率的摄像机、高速数据采集卡、工业PC机以及相关软件组成。分析了测量中相关的误差,并提出了相应措施。在整个测量方案中,所用技术比较成熟,各种设备造价适中,有利于在工业现场实现和推广。
Micro-motion parallel manipulators ＇ work spaces are small, and the installation of sensors for the detection of location and posture is very difficult. To overcome this adverse condition, a detection method is contrived based on binocular vision principle. Utilizing corresponding relationship between image coordinate and world coordinate, the necessary information of feature point is extracted from two corresponding images. A model for detecting the location and posture of 3-DOF micro-motion worktable is established based on feature point matching. The method of solving location and posture has such merits as small information content, small computation, high processing speed, less time consuming, and so on. So, the method is fit for dynamic measurement. The measure system is composed of two high-resolution vidicons, high-speed data acquisition card, industrial PC and the related software. The correlated error is analyzed in the process of detecting. Some countermeasures are proposed at the same time. The technologies used are comparative maturity and the prices of equipments used are moderate in the whole measure solution. These factors help to realize and popularize the solution in the industry environment.
出处
《制造技术与机床》
CSCD
北大核心
2010年第3期109-111,共3页
Manufacturing Technology & Machine Tool
基金
江西省自然科学基金(2008GZC0051)
关键词
平面并联机器人
双目视觉
位姿
检测
Planar Parallel Manipulator
Binocular Vision
Location and Posture
Detect
