摘要
设计了一种基于视觉的5自由度清洗机械臂位姿检测系统。系统由C6000系列DSP、TVP5147M复合视频采集卡、高精度CCD以及太阳能板特征模拟样本组成。搭建了一个DSP数字图像处理实验平台,进行位姿检测实验,在实验过程中,根据使用目标板自身特性利用全局信息搜索目标特征的方法提取特征点;通过理论研究确定了基于单摄像机成像及成像过程中的坐标系变换理论;利用P4P问题和目标位姿求解算法确定目标板与机械臂之间的相对空间位置以及机械臂末端执行器的姿态。实验结果显示,系统能够实时检测到机械臂与目标工件之间的相对空间位置和姿态,当检测距离为0～800 mm,位置误差精度可以达到2.7%。
A vision-based five-degree-of-freedom cleaning robot arm pose detection system is designed.The system consists of C6000 series DSP,TVP5147M composite video capture card,high precision CCD and solar panel feature simulation samples.A DSP digital image processing experimental platform to carry out the pose detection experiment is built in the system.In the experiment process,the feature points were extracted by using the global information search target feature according to the characteristics of the target board.The theoretical research based on single camera imaging and the coordinate system transformation theory in the imaging process;the problem and the target pose detection algorithm are used to determine the relative spatial position between the target plate and the robot arm and the attitude of the end effector of the robot arm.The experimental results show that the system can detect the relative spatial position and pose between the robot arm and the target workpiece in real time.When the detection distance is within 0-800 mm,the position error accuracy can be up to 2.7%.
作者
付俊美
沈永良
王一光
FU Jun-Mei;SHEN Yong-Liang;WANG Yi-Guang(School of Electronic Engineering,Heilongjiang University,Harbin 150080,China)
出处
《黑龙江大学工程学报》
2019年第1期59-66,共8页
Journal of Engineering of Heilongjiang University
基金
黑龙江大学研究生创新科研资金项目(YJSCX2018-149HLJU)
