光学学报, 2011, 31 (4): 0415001, 网络出版: 2011-03-24   

基于神经网络的大范围空间标定技术

Large-Scale Camera Calibration with Neural Network
作者单位
国防科学技术大学机电工程与自动化学院, 湖南 长沙 410073
摘要
目前利用神经网络进行摄像机标定的研究,只能实现单个位置上或小范围空间的标定精度要求。当标定空间扩大到一定范围,标定精度和标定速度的冲突便不可调节。首先推导了摄像机的成像原理,证明沿世界坐标系3个坐标轴(XW轴,YW轴和ZW轴)方向存在截然不同的成像规律,提出沿上述3个方向分别进行标定的并行标定方法;进而提出一种新的归一化方法,较好的提高了XW轴和YW轴方向的标定精度。实验结果表明,ZW轴向标定是整个大范围标定的关键,其重构标准差远大于XW轴和YW轴的重构标准差。在保证精度与速度的前提下,新的归一化方法扩大了摄像机标定范围。
Abstract
Current researches on camera calibration using neural network can only keep precision within a small space, unfortunabely, there has no such a method that can solve conflict between calibration precision and speed as space extends. It is proved that there are different imaging rules along three axes (XW axis, YW axis and ZW axis) in the world coordinate system, and a brand new parallel calibration method is created to calibrate camera along these three directions respectively, namely parallel calibration. A new normalization method is proposed and better calibration precisions are gained along both XW axis and YW axis. There comes conclusion that calibration along ZW axis is critical to the whole large-scale calibration since standard deviation of reconstruction along ZW axis is much higher than those along XW axis and YW axis. Expersmental results show that the new normalization method extends calibration scale while keeping both calibration precision and speed.

田震, 张玘, 熊九龙, 王国超. 基于神经网络的大范围空间标定技术[J]. 光学学报, 2011, 31(4): 0415001. Tian Zhen, Zhang Qi, Xiong Jiulong, Wang Guochao. Large-Scale Camera Calibration with Neural Network[J]. Acta Optica Sinica, 2011, 31(4): 0415001.

本文已被 10 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!