Computer Vision Aided Target Linked Radiation Imaging
Abstract
In this paper, we demonstrated an application of video tracking to radiation detection, where a vision-based tracking system enables a traditional CZT (cadmium zinc telluride)-based radiation imaging device to detect radioactive targets that are in motion. An integrated real-time system consisting of multiple fixed cameras and radiation detectors was implemented and tested. The multi-camera tracking system combines multiple feature cues (such as silhouette, appearance, and geometry) from different viewing angles to ensure consistent target identities under challenging tracking conditions. Experimental results show that both the video tracking and the integrated systems perform accurately and persistently under various scenarios involving multiple vehicles, driving speeds, and driving patterns. The results also validate and reiterate the importance of video tracking as an enabling technology in the field of radiation imaging.
Cite
Text
Gao et al. "Computer Vision Aided Target Linked Radiation Imaging." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012. doi:10.1109/CVPR.2012.6247797Markdown
[Gao et al. "Computer Vision Aided Target Linked Radiation Imaging." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012.](https://mlanthology.org/cvpr/2012/gao2012cvpr-computer/) doi:10.1109/CVPR.2012.6247797BibTeX
@inproceedings{gao2012cvpr-computer,
title = {{Computer Vision Aided Target Linked Radiation Imaging}},
author = {Gao, Dashan and Yao, Yi and Pan, Feng and Yu, Ting and Yu, Bing and Guan, Li and Dixon, Walter and Yanoff, Brian and Tian, Tai-Peng and Krahnstoever, Nils},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2012},
pages = {1162-1169},
doi = {10.1109/CVPR.2012.6247797},
url = {https://mlanthology.org/cvpr/2012/gao2012cvpr-computer/}
}