Human Height Measurement in Surveillance Video Based on Vision Technology

Fei Yuan ； Junji He

《International Core Journal of Engineering》 6卷5期 (2020/05) Pp. 198-208

https://doi.org/10.6919/ICJE.202005_6(5).0030

摘要

Traditional human height measurement technology in surveillance video usually requires a three-dimensional scene reconstructed which combined with the camera calibration to get an accurate measurement. Or sometimes, the approximate height of the human body is determined based on the principle of projective geometry without calibration. In this paper, a new visual measurement method is proposed which can accurately measure the height of a human body in a surveillance video without reconstructing a three-dimensional scene. In the case of camera calibration, this method uses the principle of pinhole imaging, and deduces the homography matrix of the human body plane through the distance between the human body plane and the reference wall, and then establishes the human height model. Moreover, in the implementation process, the corner information in the calibration target is fully used to construct the horizontal vanishing point and the vertical vanishing point in space. Then, the overhead points and the perpendicular points needed in the human height measurement model are extracted effectively. The experimental results show that the error of the measurement result can be less than 1.28%, which can meet the needs of human height measurement.

關鍵字

Camera calibration ； Surveillance video ； Vanishing point ； Homography matrix ； Height measurement

參考文獻

Criminisi A. Accurate Visual Metrology from Single and Multiple Uncalibrated Images[M]. Springer Press,2002.

Qiulei Dong, Yihong Wu, Zhanyi Hu. Video-based Real-time Automatic Human Height Measurement[J]. Acta Automatica Sinica, 2009, 35(02):137-144.

Zhang Zhengyou. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11):1330-1334.

Zhijie Gan, Yun Liu. Real-time Stature Measurement Algorithm Based on Monocular Vision Technique [J]. Journal of Qingdao University of Science and Technology (Natural Science Edition), 2008, 029(004):366-369.

Caixia Zhang, Huanli Fu. Visual Metrology for Height of Pedestrian from Uncalibrated Video[J]. Computer Engineering and Applications, 2017, 53(21):162-166.