相机标定的金字塔迭代优化算法

doi:10.3778/j.issn.1002-8331.2103-0347

摘要/Abstract

摘要： 在当前交通场景中，位姿、焦距均可变的云台相机正在被大规模应用，现有的相机标定模型在求解相机参数的过程中，通常使用先验条件一次性求解相机参数的方法。这种方法在云台相机不断改变位姿和焦距的过程中，标定结果存在不稳定、不精确的情况，其中相机高度的不稳定现象尤为明显。针对这个问题，提出一种基于焦距和相机高度的金字塔迭代优化方案。使用基础的相机标定，得到一个基础标定结果；设定沿道路方向优化对比内容为道路标线，垂直于道路方向上为道路宽度；设定焦距和相机高度优化范围及初始优化步长，并采用金字塔迭代的方式进行结果优化，最终得到最佳优化结果。在高速公路的数据集上的实验结果表明，该方法能够有效提高相机标定参数的精度和稳定性，比之传统标定算法，在沿道路方向上精度提高8%，垂直于道路方向精度提高6.5%，综合精度提高6.5%；在云台相机改变位姿和焦距过程中，相机高度波动稳定在2%以内。

关键词: 相机标定, 云台相机, 金字塔迭代优化, 交通场景, 道路标志物

Abstract: In current traffic scenes, pan-tilt cameras with variable pose and focal length are being used on a large scale. In the process of solving camera parameters, the existing camera calibration model usually uses the method which uses the prior conditions in the scene to solve the camera at one time. In this method, the calibration results are unstable and inaccurate when the PTZ camera continuously changes the pose and focal length. Among them, the unstable phenomenon of the camera height is particularly obvious. To solve the problem, this paper proposes a pyramid iterative optimization scheme based on focal length and camera height. First it uses the basic camera calibration to obtain a basic calibration result; it sets the optimized comparison content along the road as road markings, and the comparison content is perpendi-
cular to the road as the road width; it sets the focal length and camera height optimization range and initial optimization step length, and uses the pyramid iteration method to optimize the results, finally gets the best optimization results. The experimental results on the highway data set show that this method can effectively improve the accuracy and stability of the camera calibration parameters. Compared with the traditional calibration algorithm, the accuracy in the direction along the road is increased by 8%, and the accuracy in the direction perpendicular to the road is increased by 6.5%, and the overall accuracy is increased by 6.5%; during the process of changing the pose and focal length of the pan-tilt camera, the camera height fluctuation stabilizes within 2%.

Key words: camera calibration, pan-tilt-zoom camera, pyramid iterative optimization, traffic scene, road prior knowledge

李聪亮, 宋焕生, 穆勃辰, 张文涛. 相机标定的金字塔迭代优化算法[J]. 计算机工程与应用, 2022, 58(20): 286-292.

LI Congliang, SONG Huansheng, MU Bochen, ZHANG Wentao. Pyramid Iterative Optimization Algorithm for Camera Calibration[J]. Computer Engineering and Applications, 2022, 58(20): 286-292.

参考文献

[1] XU Y，YU G，WANG X，et al.A hybrid vehicle detection method based on Viola-Jones and HOG+SVM from UAV images[J].Sensors，2016，16（8）：1325.
[2] SOCHOR J，JURANEK R，SPANHEL J，et al.Comprehensive data set for automatic single camera visual speed measurement[J].IEEE Transactions on Intelligent Transportation Systems，2018：1-11.
[3] SU Y，CUI X，XI Q，et al.An industry-ready single PTZ-camera based attendance management system[C]//2020 IEEE International Conference on Multimedia & Expo Workshops（ICMEW），2020.
[4] FANG Y，WU J J，HUANG B M.2D sparse signal recovery via 2D orthogonal matching pursuit[J].Science China Information Sciences，2012，55（4）：889-897.
[5] SOCHOR J，JURANEK R，HEROUT A.Traffic surveillance camera calibration by 3D model bounding box alignment for accurate vehicle speed measurement[J].Computer Vision & Image Understanding，2017，161：87-98.
[6] CHEN J，LITTLR J J.Sports camera calibration via synthetic data[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops，2019.
[7] BARTL V，?PANHEL J，DOBES P，et al.Automatic camera calibration by landmarks on rigid objects[J].Machine Vision and Applications，2021，32（1）：1-13.
[8] ZHANG Z.A flexible new technique for camera calibration[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2000，22（11）：1330-1334.
[9] 王娜.基于运动车辆的摄像机自动标定研究[D].北京：北京邮电大学，2019.
WANG N.Research on automatic camera calibration based on moving vehicles[D].Beijing：Beijing University of Posts and Telecommunications，2019.
[10] KANHERE N K，BIRCHFIELD S T.A taxonomy and analysis of camera calibration methods for traffic monitoring applications[J].IEEE Transactions on Intelligent Transportation Systems，2010，11（2）：441-452.
[11] SOCHOR J，JURANEK R，?PANHEL J，et al.Comprehensive data set for automatic single camera visual speed measurement[J].IEEE Transactions on Intelligent Transportation Systems，2018，20（5）：1633-1643.
[12] ZHENG Y，PENG S.A practical roadside camera calibration method based on least squares optimization[J].IEEE Transactions on Intelligent Transportation Systems，2014，15（2）：831-843.
[13] BHARDWAJ R，TUMMALA G K，RAMALINGAM G，et al.Autocalib：automatic traffic camera calibration at scale[C]//Proceedings of the 4th ACM International Conference on Systems for Energy-Efficient Built Environments.New York：ACM Press，2017.
[14] 朱忆.交通场景中基于运动信息的摄像机标定实现[D].南京：东南大学，2012.
ZHU Y.Implementation of camera calibration based on motion information in traffic scenes[D].Nanjing：Southeast University，2012.
[15] CHUANG J H，HO C H，UMAM A，et al.Geometry-based camera calibration using closed-form solution of principal line[J].IEEE Transactions on Image Processing，2021，30：2599-2610.
[16] XU Y，GAO F，REN H，et al.An iterative distortion compensation algorithm for camera calibration based on phase target[J].Sensors，2017，17（6）：1188.
[17] 郑宝峰，张朝阳，张向清，等.交通场景中的相机标定方法研究[J].科技风，2018（11）：11.
ZHENG B F，ZHANG C Y，ZHANG X Q，et al.Research on camera calibration methods in traffic scenes[J].Technology Wind，2018，（11）：11.
[18] 杨风开，杨红亮，程素霞.双摄像机模组的组合式标定和校正方法[J].计算机工程与科学，2019，41（2）：302-307.
YANG F K，YANG H L，CHENG S X.A combined calibration and correction method for dual-camera module[J].Computer Engineering and Science，2019，41（2）：302-307.
[19] 武非凡，梁浩翔，宋焕生，等.基于相机标定的跨相机场景拼接方法[J].计算机系统应用，2020，29（1）：176-183.
WU F F，LIANG H X，SONG H S，et al.Multi-camera traffic scene mosaic based on camera calibration[J].Computer Systems & Applications，2020，29（1）：176-183.