A fully automated calibration method for an optical see-through head-mounted operating microscope with variable zoom and focus

A fully automated calibration method for an optical see-through head-mounted operating microscope with variable zoom and focus
复制标题

DOI:
10.1109/tmi.2005.856746
复制
发表时间:
2005-11-01
影响因子:
10.6
通讯作者:
Birkfellner, W
Birkfellner, W
中科院分区:
工程技术1区
文献类型:
--
作者:
Figl, M;Ede, C;Birkfellner, W

文献摘要

被引文献

相似文献

Ever since the development of the first applications in image-guided therapy (IGT), the use of head-mounted displays (HMDs) was considered an important extension of existing IGT technologies. Several approaches to utilizing HMDs and modified medical devices for augmented reality (AR) visualization were implemented. These approaches include video-see through systems, semitransparent mirrors, modified endoscopes, and modified operating microscopes. Common to all these devices is the fact that a precise calibration between the display and three-dimensional coordinates in the patient's frame of reference is compulsory. In optical see-through devices based on complex optical systems such as operating microscopes or operating binoculars-as in the case of the system presented in this paper-this procedure can become increasingly difficult since precise camera calibration for every focus and zoom position is required. We present a method for fully automatic calibration of the operating binocular VariosCope (TM) M5 AR for the full range of zoom and focus settings available. Our method uses a special calibration pattern, a linear guide driven by a stepping motor, and special calibration software. The overlay error in the calibration plane was found to be 0.14-0.91 mm, which is less than 1% of the field of view. Using the motorized calibration rig as presented in the paper, we were also able to assess the dynamic latency when viewing augmentation graphics on a mobile target; spatial displacement due to latency was found to be in the range of 1.1-2.8 mm maximum, the disparity between the true object and its computed overlay represented latency of 0.1 s. We conclude that the automatic calibration method presented in this paper is sufficient in terms of accuracy and time requirements for standard uses of optical see-through systems in a clinical environment.