基于机器视觉的空间组合体地面实验位姿测量技术研究

Position and pose measurement technology of space combination ground experiment based on machine vision

针对空间组合体如绳系系统在地面模拟实验时具有变化复杂和响应快等特点难以实时测量问题,对空间组合体地面模拟实验系统中物体位姿测量和控制方法进行了研究。将非接触式测量的机器视觉技术应用到了组合体地面测量系统中,并从减少图像搜索位置和每个位置处的计算量角度,提出了一种改进的多模板匹配算法配合刚体平面运动位姿算法组成测量系统,以实现运动目标物体图像的实时采集;获取了目标物体上标志点的坐标,再通过坐标变换等计算动态得到了平面刚体运动位姿;利用所研制的地面模拟检测系统进行了直线运动、旋转运动和刚体平面运动等实验测试。研究结果表明:位姿测量系统的最大测量误差不超过5%,采样时间为0.2 s,可满足绳系组合体等空间组合体的测量和控制要求。

Aiming at the problem that space combination,such as the tethered system can not be measured in real time because of the characteristics of complex and fast response in spatial operations. The measurement and control methods of object position and pose in the ground simulation experiment system were studied. The machine vision technology of non-contact measurement was applied to the ground measurement system. An improved multiple template matching algorithm was proposed to constitute a measurement system combined with rigid plane motion position and pose algorithm in the perspective of reducing the computational complexity of the image search location and each location.The moving target images and the coordinates of the targets on the markers were dynamically acquired. The motion position and pose of the plane rigid body was calculated in real time by coordinate transformation. The experiment of linear motion,rotation motion and rigid plane motion were carried out by using the developed ground simulation detection system. The results indicate that the maximum measurement error of the position and pose measurement system is less than 5%,and the sampling time is 0. 2 s,which can meet the requirements of measurement and control of space tethered combination and other combinations.