基于多视觉线结构光传感器的大尺度测量方法

发布时间：2018-05-29 18:35

本文选题：测量 + 大尺度测量　；参考：《中国激光》2017年11期

【摘要】：为解决现有线结构光大尺度测量标定复杂、精度不高的问题,提出了一种基于多视觉线结构光传感器的大尺度、高精度测量方法。多视觉线结构光传感器含1个激光器和多个等间距并排的相机(一主多从),相互间有1/3左右重叠视场(OFOV),光束覆盖所有相机视宽。传感器标定只需简易的棋盘格靶标,按张氏标定法采集靶标图像、标定传统参数;此外,考虑相机视角的不变性,利用OFOV内标定图像的已知特征角点,经图像配准可预标定相邻相机图像拼接的变换矩阵。变换矩阵的逐步连乘得到任意从相机图像转换到主相机成像平面的透视变换模型(PTM)。该方法通过各相机同步采集大尺度物体的局部光条图像,再利用PTM将所有局部光条快速拼接成完整的光条图像,最终经光条坐标提取、换算得到光条位置的三维坐标。实验结果表明:与已有方法相比,该方法使用简便、精度更高,重构模型平均构造深度与真实模型仅差8.3%。
[Abstract]:In order to solve the problem of complex calibration and low precision of the existing line structure light measurement, a large scale and high precision measurement method based on multi vision line structure light sensor is proposed. The multi-vision-line structured light sensor consists of a laser and a plurality of cameras with equal spacing (one master and more than one follower), with about one third of each other overlapping field of view (OFOV), and the beam covers all the camera's viewing width. Sensor calibration only needs simple chessboard grid target, according to Zhang's calibration method to collect target image and calibrate traditional parameters. In addition, considering the invariance of camera angle, the known feature corner of image is calibrated in OFOV. The image registration can precalibrate the transform matrix of the adjacent camera image stitching. The model of perspective transformation from camera image to the imaging plane of the main camera is obtained by stepwise multiplication of the transformation matrix. In this method, the local light strip images of large scale objects are collected synchronously by each camera, and then all the local light strips are quickly spliced into complete optical strip images by PTM. Finally, the three dimensional coordinates of the position of the light strip are obtained by extracting the optical strip coordinates. The experimental results show that the method is simpler and more accurate than the existing methods, and the average construction depth of the reconstructed model is only 8.3% less than that of the real model.
【作者单位】：中国矿业大学(北京)煤炭资源与安全开采国家重点实验室;萍乡学院机械电子工程学院;
【基金】：国家自然科学基金(61601466) 煤炭资源与安全开采国家重点实验室开放课题(SKLCRSM16KFD03)
【分类号】：TP212;TP391.41

【相似文献】