基于改进野犬优化算法的太赫兹线阵相机内参优化  

Internal parameter optimization of terahertz linear scan camera based on improved Dingo optimization algorithm

作　　者：方灿 方波 蔡晋辉 FANG Can;FANG Bo;CAI Jinhui(College of Metrology and Measurement Engineering,China Jiliang University,Hangzhou 330018,China)

机构地区：[1]中国计量大学计量测试工程学院,浙江杭州330018

出　　处：《量子电子学报》2025年第2期157-164,共8页Chinese Journal of Quantum Electronics

基　　金：浙江省自然科学基金项目(LY22F050001);浙江省属高校基本科研业务费专项资金(2021YW09)。

摘　　要：太赫兹透射式成像具有对人体无害、对非金属材料的穿透性较强等特点,因此可用于安全检验、生物医学等领域。提高太赫兹线阵相机测量精度的关键在于相机参数的准确性。本文提出一种改进的野犬优化算法(DOA),对Draréni标定法所得的相机参数进行优化。该算法改良了DOA中的食腐和生存行为,并引入粒子群优化算法和灰狼算法的个体更新策略来增强全局搜索能力。利用自行搭建的太赫兹透射式扫描成像装置进行成像,选取10张不同位置的成像图片作为测试对象,并利用Draréni标定法得到相机初始内外参数,最后,分别利用粒子群优化算法、DOA和本文提出的改进的DOA进行相机内参优化。实验结果表明,本文提出的改进DOA相对于传统Draréni标定法、粒子群优化算法和标准的DOA,其平均重投影误差分别降低了33.41%、21.35%和12.62%,证实了该算法具有较高的稳定性,并能显著提高相机标定的精度。Terchertz(THz)transmission imaging is characterized by its harmless nature to human body and high penetration capability for non-metallic materials,making it suitable for applications in security inspection and biomedical fields.The key to improve the measurement accuracy of THz line scan camera lies in the accuracy of camera parameters.This paper proposes an improved Dingo optimization algorithm(DOA)to optimize the camera parameters obtained by the Draréni calibration method.The improved DOA enhances the global search capability by modifying the feeding and survival behaviors in the traditional DOA,and introducing the individual updating strategy of particle swarm optimization algorithm and grey wolf algorithm.A THz transmission scanning imaging device is built for imaging,and 10 images from different positions are selected as test objects.The camera's initial intrinsic and extrinsic parameters are obtained using the Draréni calibration method.Finally,the particle swarm optimization algorithm,DOA,and the improved DOA proposed in this paper are used respectively for camera internal parameter optimization.The experimental results show that the proposed algorithm in this paper reduces the average reprojection error by 33.41%,21.35%,12.62%respectively compared with the traditional Draréni calibration method,the particle swarm optimization algorithm,and the standard DOA,which demonstrates that the proposed algorithm is stable and can significantly improve the accuracy of camera calibration.

关 键 词：图像与信息处理 太赫兹成像 线阵相机标定 参数优化 改进的野犬优化算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]