面向扑翼飞行机器人的电子稳像算法设计  

Design of an electronic image stabilization algorithm for flapping-wing flying robots

作　　者：刘胜南 付强[1,2,3] 冯楠 张春华 贺威[1,2,3] LIU Shengnan;FU Qiang;FENG Nan;ZHANG Chunhua;HE Wei(School of Intelligence Science and Technology,University of Science and Technology Beijing,Beijing 100083,China;Institute of Artificial Intelligence,University of Science and Technology Beijing,Beijing 100083,China;Key Laboratory of Intelligent Bionic Unmanned Systems(Ministry of Education),University of Science and Technology Beijing,Beijing 100083,China;Automation Research Institute Co.,Ltd.,China South Industries Group Corporation,Mianyang 621000,China)

机构地区：[1]北京科技大学智能科学与技术学院,北京100083 [2]北京科技大学人工智能研究院,北京100083 [3]北京科技大学智能仿生无人系统教育部重点实验室,北京100083 [4]中国兵器装备集团自动化研究所有限公司,绵阳621000

出　　处：《工程科学学报》2024年第9期1544-1553,共10页Chinese Journal of Engineering

基　　金：科技部雄安新区科技创新专项项目(2023XAGG0062);国家自然科学基金资助项目(62173031,U20A20225);北京科技大学青年教师学科交叉研究资助项目(FRF-IDRY-22-029);北京科技大学中央高校基本科研业务费专项资金资助项目(FRF-TP-22-003C2)。

摘　　要：在扑翼飞行机器人的飞行过程中,由于其特有的驱动方式,机身存在周期性的俯仰和滚转运动,导致航拍视频出现抖动,影响成像效果的清晰度和稳定性.为了解决这一问题,本文提出了一种基于ORB和滑动均值滤波的电子稳像算法,用于在线消抖处理.首先,针对扑翼飞行机器人航拍图像的抖动周期与机翼扑动周期相一致的特性,设计了一种估计算法来根据图像特征估计机翼扑动周期.这一算法能够更准确地捕捉到抖动的周期性特征,为后续的稳像处理提供了重要参数.其次,提出了一种与扑动周期相关联的运动滤波算法,能够根据不同飞行工况对滤波参数自适应地进行动态调整.本文提出的算法优点在于能够根据扑翼飞行机器人实际飞行情况实时调整参数,从而更好地适应不同的飞行工况,进一步提高了稳像效果.最后,为了验证算法的可行性和稳定性,将视觉成像装置搭载在扑翼飞行机器人上进行了飞行实验.实验结果表明,相较于常用的电子稳像算法,本文所设计的算法在扑翼飞行机器人中表现出更好的稳像效果.最后,总结了本文所提出的算法优点,并对未来研究方向做出了展望.During the flight of a flapping-wing flying robot,the unique flapping-wing propulsion mechanism causes periodic pitching and rolling motions of the body.In passing through different airspeeds and altitudes over different terrain,the wings cycle upand-down through predefined stroke patterns to generate the aerodynamic force required for powered flight.However,this oscillatory flapping motion also causes the aircraft structure to pitch and roll periodically about its center of mass in an oscillatory manner.Consequently,substantial high-frequency jitter shakes aerial video footage captured by the onboard optical sensors.In particular,the rapid shaking that disrupts the image is synchronized with the characteristic wing beat rhythm.This jitter negatively affects the quality and usefulness of the acquired aerial video.Repeated up-and-down pitching and rolling displacements shake aerial footage,greatly reducing its clarity and stability.Without mitigation,the jitter severely affects imaging results,limiting the potential applications of such video.To solve this problem,this paper proposes an electronic image stabilization algorithm based on oriented fast and rotated brief(ORB)and sliding mean filtering for online debounce processing.First,because the jitter period of aerial images of a flapping-wing flying robot is consistent with the wing flapping period,we designed an estimation algorithm to estimate the wing flapping period based on image features.This algorithm enables us to more accurately capture the periodic characteristics of jitter and provides important parameters for subsequent image stabilization processing.Second,we proposed a motion filtering algorithm associated with the flapping period,which can adaptively and dynamically adjust filtering parameters adaptively according to different flight conditions.The algorithm proposed in this paper is advantageous because it can dynamically adjust parameters in real time based on the actual flight conditions of flapping-wing robots,thereby better adapting to differ

关 键 词：扑翼飞行机器人 电子稳像算法 ORB算法 滑动均值滤波 扑动周期 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]