【24h】

LADYBUG3 CAMERA CALIBRATION

机译:LADYBUG3相机校准

获取原文

摘要

The applications of photogrammetry such as aerial triangulation, 3D modeling, and automatic texture mapping need precise camera internal geometry in order to obtain high precision results. Ladybug3 equipped with 6 short focal length fish-eye cameras, which can acquire 720 degrees panoramic image and is useful for augmented reality applications. The field of view (FOV) of each lens is about 95 degrees by 77 degrees. Thus, their lens distortion is serious and the way to calibrate their interior orientation parameters is critical for photogrammetry purposes. Therefore, this research proposes two types of indoor calibration field to solve these problems using single camera calibration and multiple camera calibration. A rotatable calibration field contains 112 pillars with Australis© coded target is designed. The basic idea of such calibration field is to acquire high convergent imaging geometry with lots of redundant image measurements by means of automatic code target recognition. The purpose of the second calibration field is to calibrate both the interior orientation parameters and relative orientation parameters between the six cameras. A 3D cube, with 4 meters in length, 3 meters in width and 2.5 meters in height is used, where 118 Australis coded targets are fixed and spread uniformly to the ceiling, the floor, and all four walls. Since the Australis is capable of self-calibrating multiple cameras at the same time, the interior and exterior orientation parameters can be solved. Based on the exterior orientation parameters, the relative orientation parameters can be estimated. The accuracy of relative rotation angles is around 0.02 degrees and the accuracy of perspective center offset which between the referenced camera to the others is less than 1 mm. It is expected that the accuracy is precise enough for direct georeferencing applications using a land vehicle mobile mapping system.
机译:摄影测量学的应用,例如空中三角测量,3D建模和自动纹理映射,需要精确的相机内部几何形状才能获得高精度结果。 Ladybug3配备6个短焦距鱼眼镜头,可以获取720度全景图像,对于增强现实应用很有用。每个镜头的视场(FOV)约为95度乘77度。因此,它们的透镜畸变是严重的,并且校准它们的内部取向参数的方法对于摄影测量目的是至关重要的。因此,本研究提出了两种类型的室内标定场来解决使用单摄像机标定和多摄像机标定的这些问题。一个可旋转的校准区域包含112个带有Australis©编码目标的柱子。这种校准领域的基本思想是通过自动代码目标识别来获取具有大量冗余图像测量结果的高会聚成像几何。第二校准字段的目的是校准六个相机之间的内部方向参数和相对方向参数。使用一个3D立方体,该立方体的长度为4米,宽度为3米,高度为2.5米,其中固定了118个Australis编码目标,并均匀地分布在天花板,地板和所有四个墙壁上。由于Australis能够同时自校准多台摄像机,因此可以解决内部和外部方向参数。基于外部取向参数,可以估计相对取向参数。相对旋转角度的精度约为0.02度,并且参考相机与其他相机之间的透视中心偏移的精度小于1 mm。期望对于使用陆地车辆移动测绘系统的直接地理配准应用而言,该精度足够精确。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

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