«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年03期

页码:

146-151

栏目:

人工智能

出版日期:

2023-03-10

文章信息/Info

Title:

Camera Calibration Optimization Method Based on Improved Sparrow Search Algorithm

文章编号:

1673-629X(2023)03-01146-06

作者:

商 海; 倪受东; 苏智勇

南京工业大学 机械与动力工程学院,江苏 南京 211816

Author(s):

SHANG Hai; NI Shou-dong; SU Zhi-yong

School of Mechanical and Power Engineering,Nanjing Tech University,Nanjing 211816,China

关键词:

摄像机标定; 内外部参数; 畸变系数; 鸟群算法; 适应度函数; 平均重投影误差; 改进麻雀搜索算法

Keywords:

camera calibration; internal and external parameters; distortion coefficient; bird swarm algorithm; fitness function; average re-projection error; improved sparrow search algorithm

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 03. 022

摘要:

目前一些摄像机标定方法,比如 DLT 标定法、Tsai 标定法和张正友标定法,都有着简单且易标定的优点,但也存在着标定精度低、鲁棒性差等缺点。 为了有效地解决这些问题,在麻雀搜索算法的基础上考虑其与鸟群算法相结合,提出了一种改进麻雀搜索算法　( improved sparrow search algorithm,ISSA)　对目标摄像机进行标定。 首先,利用 MATLAB 对标定板进行预拍摄;其次,利用 MATLAB 软件中自带的标定工具箱对采集的图像进行预标定,得到初始的摄像机内外参数;然后,构建平均重投影误差的适应度函数,并用 ISSA 对构建的平均重投影误差的适应度函数进行优化,利用适应度函数的优化对内外部参数进行优化;最后,与基于麻雀搜索算法、天牛须搜索算法( BAS) 的摄像机标定方法进行实验对比,发现基于 ISSA、SSA 和 BAS 的摄像机标定方法的平均重投影误差分别为 0. 002 9 pixel、0. 004 9 pixel 和 0. 003 7 pixel,说明ISSA 算法相对于另外两个算法在标定精度上有着一定的提升,且稳定性与鲁棒性都有所提高。

Abstract:

At present,some camera calibration methods,such as DLT calibration,Tsai calibration and Zhang Zhengyou calibration,havethe advantages of simple and easy calibration,but also have the disadvantages of low calibration accuracy,poor robustness and so on. Inorder to effectively solve these problems,an improved sparrow search algorithm ( ISSA) is proposed to calibrate the target camera basedon the combination of sparrow search algorithm and bird swarm algorithm. Firstly, the calibration board is pre - photographed byMATLAB. Secondly,the collected images are pre - calibrated by using the calibration toolbox in MATLAB software, and the initialcamera internal and external parameters are obtained. Then, the fitness function of the average re - projection error is constructed andoptimized by ISSA,and the internal and external parameters are optimized by the optimization of the fitness function. Finally,comparedwith the camera calibration methods based on sparrow search algorithm and beetle antennae search algorithm ( BAS) ,it is found that theaverage re-projection errors of the camera calibration methods based on ISSA,SSA and BAS are 0. 002 9 pixel,0. 004 9 pixel and 0. 003 7pixel respectively,which shows that the calibration accuracy of ISSA algorithm is improved compared with the other two algorithms,andthe stability and robustness are improved.

常用功能

更新日期/Last Update: 2023-03-10