基于Meanshift与摄像机标定的无人机定位方法  

Positioning method of unmanned aerial vehicles based on Meanshift and camera calibration

作　　者：成怡[1] 王晓红[1] 修春波[1] CHENG Yi;WANG Xiao-hong;XIU Chun-bo(School of Electronic Engineering and Automation, Tianjin Polytechnic University, Tianjin 300387, China)

机构地区：[1]天津工业大学电气工程与自动化学院,天津300387

出　　处：《天津工业大学学报》2017年第1期55-60,共6页Journal of Tiangong University

基　　金：国家自然科学基金资助项目(61203302);天津市应用基础与前沿技术研究计划项目(14JCYBJC18900)

摘　　要：针对在全球定位系统信号失效的情况下,提出一种利用机载相机拍摄的移动物体轨迹及其与无人机的相对位置解算无人机的绝对位置的方法.将Meanshift目标跟踪算法与摄像机标定技术相融合进行位置解算,即根据获得的摄像机内外参数、移动物体的图像像素坐标和无人机与移动物体的相对位置确定无人机的世界坐标.经多组实验数据验证,位置解算值绝对误差均不超过真实值的1%,分析结果证明了本方法在无人机自主定位领域的可行性.Under the condition of the signal failure of the global positioning system(GPS), a method for calculating absolute position of unmanned aerial vehicles(UAV) by using moving targets locus filmed by onboard camera and the relative position between moving targets and UAV is proposed. This method combines Meanshift target tracking algorithm with camera calibration technology to calculate the position. According to the obtained internal and external camera parameters, image pixel coordinates of moving targets and relative position between UAV and moving targets, the world coordinates of UAV can be calculated. Multiple sets of experimental data verifies that the absolute error of calculating data is no more than 1% of the real data. Simulation results validate the feasibility of the proposed method in the field of UAV autonomous flight.

关 键 词：摄像机标定 无人机 定位 目标追踪 MEANSHIFT算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]