【作者】 冯现东；

【导师】 王宪伦；

【作者基本信息】 青岛科技大学 ， 机械工程（专业学位）， 2018， 硕士

【摘要】 “人类对飞行的梦想是与生俱来的。”各类飞行器应运而生。按照飞行器的螺旋桨布局结构分类,飞行器可以分为固定翼、直升机、多旋翼等类型。近些年来,旋翼类飞行器在影视航拍、高校科研、救灾灭火、农林畜牧业、国防防务安全等众多高新科技领域快速发展,其独特的布局和使用方式使其在应用过程当中具有其他类型飞行器无法替代的优势。其中,无人机室内避障技术的研究在保护室内人员的安全及无人机的安全具有重要的研究价值,而场景3D重建技术需要依靠飞行器搭载相机拍摄的图像画面,真实地还原出无人机所处的三维空间场景,对航空航天、古迹建筑三维重建、增强现实(AR)技术等诸多领域具有重要意义。本文介绍了国内外无人机室内躲避障碍物以及场景三维重建技术的研究发展现状,设计了一种新型四旋翼无人机,并以此作为实验平台,研究了飞行器室内躲避障碍物及场景三维图形建模技术,主要研究内容如下。本课题首先介绍了多旋翼飞行器的分类,并介绍各类多旋翼飞行器的优缺点及主要应用领域范围,给出了四旋翼飞行器的飞行动力学原理,确定了适合本课题的设计方案,根据飞行器的基本组成要素,设计出了创新型的机架与机翼结构,并对其结构强度进行有限元分析,以满足符合飞行器碰撞强度要求。其次,给出了姿态解算算法的推导过程,采用牛顿力学法建立了飞行器数学模型,将超声波传感器合理地布局在飞行器机体周围,实现室内避障,以便能够更好地探测到障碍物存在的位置及距离。在完成无人机室内避障系统开发和搭建后,对搭载超声波避障系统的无人机进行室内避障飞行实验,并对实验产生的超声波探测的距离数据和障碍物方位进行分析,分析验证飞行器避障系统的安全性以及稳定性。最后,研究了基于摄像头的场景三维重建技术,依据计算机视觉的三维重建过程,完成Kinect摄像头参数的标定,求出摄像头内参数矩阵,介绍基于图像的二维特征提取算法SIFT和SURF,并结合理论计算分析两种特征提取法的优缺点,设计场景特征提取实验加以验证。通过实验,拍摄室内场景的二维图像,在计算机端通过三维重建技术,重建室内的场景,分析成像质量,评估算法优缺点。更多还原

【Abstract】 “Human beings are born with the dream of flying.” So all kinds of aircrafts came into being.According to the classification of aircraft propeller layout structure,the airarafts are divided into fixed wing,helicopter,multi_rotor.In recent years,multi-rotor aircrafts developed rapidly in the high-tech fields of aerophotography,scientific research,fire control,agroforestry,national defense security.The unique layout and use mode makes it an advantage that other types of aircraft cannot replace in the application process.The research of the UAV indoor obstacle avoidance technology has important research value in protecting the safety of personnel and aircraft.The 3D reconstruction technology can veritably restore the three-dimentional space scene of the images captured by the aircraft’s airborne camera,which has important significance to the aerospace,heritage building 3D reconstruction,AR.This paper introduced the development status of indoor UAV obstacle avoidance and 3D scene reconstruction technology,and designed the brand-new four-rotor aircraft as the experimental platform for the indoor aircraft obstacle avoidance and 3D scene reconstruction technology.The main research contents are as follows.This paper firstly introduced the classification of multi-rotor aircraft,and analyzed the advantages and disadvantages of all kinds of multi-rotor aircrafts as well as major application areas.The four-rotor aircraft dynamics was given to determine the suitable four-rotor structure for this topic.According to the basic components of the four-rotor aircraft,the new four-rotor aircraft was designed and the finite element analysis of the structure was carried out to meet the requirements of the collision intensity.Secondly,this paper introduced the process of the attitude calculation algorithm in detail.The mathematical model of aircraft was built using Newtonian mechanics method.Then,the ultrasonic sensor was reasonably arranged around the body of the aircraft to avoid indoor obstacles in order to better detect the obstacles’ location and distance from the aircraft.After the completion of the development and implementation of the indoor aircraft obstacle avoidance system,this paper carried out the ultrasonic obstacle avoidance system of uav indoor obstacle avoidance flight test,and analyzed the test data to verify the security and stability of obstacle avoidance system.Finally,this paper studied the 3D scene reconstruction technology of the aircraft airborne camera.Based on the principle of computer vision 3D reconstruction,this paper completed camera parameter calibration,and obtained the camera parameter matrix.SIFT and SURF were introduced based on 2D image-based feature extraction algorithm.Through the theoretical calculation,this paper analyzed the advantages and disadvantages of the two algorithms,and designed experiments to verify.Through the experiment of taking the 2D images,this paper reconstructed indoor scene and analyzed image quality,the advantages and disadvantages of the algorithm.更多还原