基于SLAM的连续运动测量坐标转换算法研究

Continuous Movement Measuring Coordinate Transformation Algorithm Based on SLAM

摘要:

近年来,视觉导航开始成为人们的研究热点;在轨道检测中,现有运动状态下的测量方法所采用的坐标系都是建立在测量系统自身上的一种连续动态变化的坐标系,不能与静态基准点建立明确的解析联系;针对这一问题,结合同步定位与地图创建的思想,提出了一种新的检测算法;基于SLAM原理,通过双目视觉系统的坐标转换方法,建立动态坐标系与静态基准点之间的关系,求得旋转平移矩阵,将后面N个位置的坐标均转换到初始坐标系下,实现连续运动测量;通过实验,设计运动路线,验证坐标转换算法,并得到了建立在世界坐标系下的空间线形曲线;实验结果表明:转换后的坐标,x、y、z三个方向的最大误差不超过0.734 mm。

Abstract:

In recent years, the visual navigation began to become the focus. In rail detection, the existing sports under the state of measuring methods adopted by the coordinate system are based on the measurement system itself a continuous dynamic changes of the coordinate system, cannot establish clear resolution contact with static reference point. In order to solve this problem, a new detection algorithm is proposed combining with simultaneous localization and mapping. In this paper, a relationship between a dynamic coordinate system and the static reference point was built to realize continuous motion measurement, which was based on the principle of simultaneous localization and mapping (SLAM) and the coordinate transformation method of binocular vision system. We designed a movement course in the experiment to verify the coordinate transformation algorithm, and got a space linear curve based on the world coordinate system. Experimental result shows that after the transformed coordinates, x, y, z three directions of the maximum error is less than 0.734 mm.