摘要
在基于视觉伺服的机器人控制中,准确跟踪动态目标的关键在于快速准确地完成连续图像中的目标辨识。本文在基于特征差异的彩色目标识别方法的基础上,应用卡尔曼滤波算法对运动目标位置进行预测,在预测的局部范围内搜索匹配目标,将全局搜索改变为局部搜索,大大减少了图像处理数据量,提高了机器人快速跟踪运动目标的实时性。
The key to track the moving object exactly is how to recognize the target of the sequential digital image quickly and accurately in the robot control system by visual servo.Based on the methods of colorful target recognition which is founded on character discrepancy,this paper makes use of the liner Kalman filter algorithm to forecast the position of the moving object,transfer the global searching scope to the local scope,and search the matching object in the forecasted local scope.The quantity of the data is reduced rapidly.The real time in tracking object of the robot is improved by applying these methods.
出处
《河南科技大学学报（自然科学版）》
CAS
2007年第3期35-38,共4页
Journal of Henan University of Science And Technology:Natural Science
基金
河南省教育厅自然科学研究计划项目(200610464031)
