基于并联机构动力学模型的标记成像波门预测

在基于双目视觉的并联机构位姿测量系统中,为了简化图像处理、提高位姿测量的实时性,提出一种标记成像波门预测方法。基于机构动力学模型和Unscented卡尔曼滤波(Unscented Kalman filter,UKF)算法实现并联机构位姿预测,并利用误差理论建立标记中心像点与并联机构位姿之间误差传递模型,进而实现标记成像波门预测。最后,通过仿真试验验证了标记成像波门预测的可行性。

Prediction for marker imaging window based on dynamic model of parallel mechanism

In order to simplify image processing and improve the real-time performance for binocular vision-based pose measurement of parallel mechanism,a method for predicting marker imaging window is proposed.It makes parallel mechanism pose prediction based on dynamic model of the mechanism and Unscented Kalman filter(UKF) algorithm,establishes error propagation model between marker center image and parallel mechanism pose according to error theory,and consequently realizes the prediction of marker imaging window.Finally,simulation experiments are conducted to verify the feasibility of the prediction for marker imaging window.