【作者】 龚建波；

【导师】 孙祥一；

【作者基本信息】 国防科学技术大学 ， 航空宇航科学与技术， 2011， 硕士

【摘要】 着陆是飞机飞行安全的一个关键环节,在降落过程中,飞机需要外测设备来辅助其安全着陆,而首要的就是要测量飞机的位置参数。基于视觉的测量方法近年来在航空航天及武器研制、试验中应用越来越多。由于飞机运动速度快、范围大,要求视觉测量设备的视场能够覆盖其运动空间,传统的双像机测量系统已经不能满足要求。对此,采用多摄像机通过视场拼接组成测量系统来完成整个运动场景的观测,实现对飞行降落全程测量,具有重要的应用价值。摄像机标定一直是计算机视觉中的一个重要课题,也是各种基于视觉测量的基础和关键步骤。本文对多摄像机测量系统的标定技术作了重点研究,提出采用预先标定同现场标定相结合的“两步”标定方法,实验证明本文方法正确有效,精度较高,满足测量要求。本文的主要工作有:1.研究多摄像机测量系统的结构特性,建立了视觉测量系统简易模型,并分析了影响视觉测量精度的因素,对视觉测量系统的结构布置具有一定指导意义。2.介绍了摄像机的聚焦和景深问题,研究了构成多像机视场拼接测量系统中的各像机的景深特性。3.重点研究了大视场远距离成像系统的标定技术,并利用本文的方法对单个像机进行了实验验证,同时还探讨了一种基于角度的摄像机标定方法,进行了仿真实验验证。4.推导了多个摄像机之间的位姿关系,提出建立参考像机坐标系的方法,在实验环境较好的情况下预先标定出各摄像机同参考像机的关系,然后在测量现场利用适量合作标志标定出参考像机的外参数,进而得到所有摄像机的外参数,有效克服了多像机现场标定的限制。5.根据测量系统中各个摄像机的物理参数特性,计算像机的视场范围,设计标定靶标和标定方案,利用前面研究的摄像机标定方法以及多像机位姿关系求解方法完成测量系统的标定工作,并进行现场测量验证,方法正确可行,精度满足要求。更多还原

【Abstract】 Landing is a key link of flight safety, aircraft need exterior trajectory measurement equipment to assist it landing safely in the process of landing, and it is first to get the aircraft’s location parameter. Vision-based measurement method is more and more applied in aerospace, weapon development and test recently. Because of high speed and long range capabilities, the visual field of vision-based measurement equipment should cover the aircraft’s moving spaces, the traditional double CCD cameras system already can’t meet the demands. Thus, use measuring system based on multiple camera by field stitching to observe the whole motion and measure the whole range of landing has important application value.Camera calibration is one of the important subject of Computer Vision, and it is also the basis and key step of kinds of vision-based measurement. This paper mainly studies calibration technology for multiple camera measuring system, a two step method combined with calibration in advance and in-situ is proposed. The experimental evidence obtained has shown that the method is correct and effective, which has high accuracy and satisfies the requirements of the measurement.The key work of the paper includes:1. The structural characteristics of multiple camera measuring system is researched, the model of vision-based measuring system is also build, and the accuracy influence factors are analyzed, which has some meaning for the structural arrangement of vision-based measuring system.2. Problem of camera focusing and depth of field is introduced, and characteristics of field depth of each camera which constitute the multiple camera measuring system is also researched.3. The dissertation mainly studies calibration technology for large field and long range imaging system, and give the experimental verification of single camera using suggested method. Also a camera calibration method based on angle is investigated and simulation experimental verification is given.4. Pose between cameras is derived in this paper, and reference camera coordinate system is established. The relationship between each camera and reference camera is calibrated in advance while the experimental environment is in a better situation, then the reference camera’s external parameter is calibrated by proper cooperative markers in measuring field, furthermore all camera’s external parameter is obtained, which effectively overcome the limitation of multiple cameras’ calibration in field.5. According to each camera’s physical parameters in measuring system, camera’s field range is calculated, then calibration targets and calibration scheme are designed, and then the calibration work of measuring system is completed using the camera calibration method on the previous study and the solving method of multiple camera’s relationship, the verification implementation in field proves that this method is feasible and meets request of precision.更多还原