首页> 外国专利> Method and device arrangement for high-precision, (semi-) automatic in-situ calibration and registration of a projector-camera system to a component

Method and device arrangement for high-precision, (semi-) automatic in-situ calibration and registration of a projector-camera system to a component

机译:用于高精度,(半)自动原位校准以及将投影机-摄像机系统配准到组件的方法和装置

摘要

Method for the highly accurate, (semi-) automatic in-situ calibration and registration of a projector-camera system to a component, comprising the following steps: - Projector and camera are positioned so that both devices point to the component and the component one preferably occupies the largest possible part of the image area of the projector and camera; The projector projects a sequence of structured light patterns onto the component; - The camera takes a picture of the so-lit component for each pattern; A rough estimate of the projector's main point and / or the camera's main point is taken, wherein the rough estimate of the projector's main point and / or the camera's main point is given at least once by a user or as a design-related value; Based on this, an initial projector-camera calibration (intrinsic: focal length, main point, extrinsic: rotation and translation) and a 3D reconstruction of the component is calculated by means of fundamental matrix decomposition; The 3D reconstruction is roughly aligned to a present virtual model of the component, in particular by predetermining or automatically determining rotation and / or translation and / or scaling of the 3D reconstruction; Iteratively calculated: a) a best fit of the 3D reconstruction to the virtual model of the component, b) an optimized projector camera calibration resulting from the best fit, preferably including a radial distortion parameter, and registration to the part, and c) a new 3D reconstruction based on the optimized calibration; - Calibration will stop when predefined Best Fit and / or Projector Camera Calibration thresholds are reached.
机译:将投影机-摄像机系统进行高精度,(半)自动原位校准和配准的方法,包括以下步骤:-放置投影机和摄像机,使两个设备都指向该组件,一个组件指向一个优选地占据投影仪和照相机的图像区域的最大可能部分;投影仪将一系列结构化的光图案投影到组件上; -相机为每种模式拍摄照明组件的照片;对投影仪的主要点和/或照相机的主要点进行粗略估计,其中,投影仪的主要点和/或照相机的主要点的粗略估计由用户至少给出一次或作为与设计有关的值给出;在此基础上,通过基本矩阵分解,计算出投影机的初始摄像头校准(本征:焦距,主点,非本征:旋转和平移)和3D重建。 3D重建大致与组件的当前虚拟模型对齐,特别是通过预定或自动确定3D重建的旋转和/或平移和/或缩放;迭代计算:a)3D重建与组件的虚拟模型的最佳拟合,b)由最佳拟合(最好包括径向变形参数)以及对零件的配准产生的优化的投影仪摄像机校准,以及c)a基于优化校准的新3D重建; -当达到预定义的最佳匹配和/或投影机摄像机校准阈值时,校准将停止。

著录项

  • 公开/公告号DE102016105405B4

    专利类型

  • 公开/公告日2017-07-20

    原文格式PDF

  • 申请/专利权人 EXTEND3D GMBH;

    申请/专利号DE201610105405

  • 发明设计人 CHRISTOPH RESCH;PETER KEITLER;

    申请日2016-03-23

  • 分类号G01B11/25;

  • 国家 DE

  • 入库时间 2022-08-21 13:22:20

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号