基于扩展Kalman滤波的轨道线形检测研究

提出一种双目视觉与陀螺仪结合的数据融合模型来计算轨道空间线形参数、并利用扩展卡尔曼滤波算法对轨道空间线形参数进行优化的方案.减小因双目视觉系统受环境变化的影响及陀螺仪随时间的累积误差,提高融合模型对轨道空间线形参数的测量精度,利用Matlab仿真出轨道空间线形.并利用机械手臂进行相应的实验,实验结果表明:双目视觉与陀螺仪结合的数据融合可有效实现轨道线型的空间测量,测量坐标在X、Y、Z三个方向上的位移误差不超过0.381 mm,优于传统的惯性或者视觉检测精度,降低了计算复杂度,具有较好实用性.

Study on the Detection of Linear Track Based on Extended Kalman Filter

A scheme which uses data fusion model combining binocular vision and gyro to calculate the geometry parameters of track space,and the extended Kalman filtering algorithm to optimize the geometry parameters,is proposed.The impact of the binocular vision system due to environmental changes and the cumulative error over time by gyroscope was reduced,the measurement precision of the fusion model was improved,and a more accurate track space geometry was simulated by Matlab.The corresponding experiments were carried out by using the mechanical arm.The experimental results show that the maximum error of the data fusion model of binocular vision and gyro is no more than 0.381 mm at the direction of X、Y、Z axis,which is better than the traditional inertia or visual inspection accuracy,reduces the computational complexity,and has good practicality.