一种拟人空间机械臂视觉伺服图像处理研究

A Personification Space Robot Arm Visual Servoing Image Processing Research

针对四自由度串并混联拟人空间机械臂本体,构建了基于运动控制卡模式的机器人软件和硬件系统。采用基于位置的控制结构,利用所求得的机械臂逆运动学方程,设计了由图像反馈和末端运动组成的视觉伺服控制系统。提出了一种基于SURF特征的静态目标识别算法,算法首先对目标图像提取SURF特征,并利用欧氏距离实现模板图像与目标图像特征点匹配,然后计算已匹配特征点的质心来获得目标的位置信息,最终实现了稳定的伺服定位。实验结果证明了系统的可靠性,并且证明了四自由度串并混联机械臂本体在实际运动的精确性和稳定性。

For the four degrees of freedom serial-parallel personification space robot arm, a new system which is based on the robot motion control card is proposed. By using the location of control structure and IKP inferential of robot arm, the system which consists of image feedback and terminal movement is designed. With the SURF, a static target identification algorithm is conducted. Firstly, the SURF features of template image and target image are extracted and matched by the Euclid distance method. Then the coordinates of object are gotten by calculating controid of features matched, and a stable servo positioning system is performed finally. The experimental results prove that the system is effective, and the movement is accurate and stable.