基于图像特征偏差的姿态调整策略研究  

Image-based Research on Posture Adjustment Strategy

作　　者：刘延杰[1] 景亚宁 隋聪颖[1] 

机构地区：[1]哈尔滨工业大学机器人技术与系统国家重点实验室,黑龙江哈尔滨150000

出　　处：《机械与电子》2014年第8期66-70,共5页Machinery & Electronics

基　　金：国家高技术研究发展计划"八六三"资助项目(2013AA040901);黑龙江省科研机构创新能力提升专项计划项目(YC13D004)

摘　　要：为了实现对泡沫柱空间姿态的调整,研究了基于双路视觉的泡沫柱空间位姿调整策略,建立了系统的运动学模型,在此基础上,提出了基于图像特征偏差的位姿调整策略。在仿真分析中,推导了基于Broyden方法的图像雅可比矩阵在线估计方程,建立了系统的运动学模型与视觉成像模型,设计采用了比例控制器。仿真结果表明,采用的图像雅可比矩阵估计方法可以很好地跟踪目标图像特征,图像特征偏差可以快速收敛到设定要求。In order to complete the task of foam’s posture adjustment,the study of space pos-ture adjustment strategy based on dual vision is carried on and the system kinematic model is estab-lished.Afterwards,a strategy based on image fea-tures deviation is proposed.In the simulation anal-ysis,the equation is derived based on image Jacobi-an matrix online estimation using the method of Broyden in which kinematic model and visual ima-ging model is established and a proportional con-troller is designed.The simulation results show that the method of image Jacobian matrix online estimation proposed in this paper can track the tar-get image well and the image feature deviation can quickly converge to the set requirements.

关 键 词：姿态调整 图像特征 图像雅可比矩阵 BROYDEN 方法 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]