一种无人机自主着陆视觉跟踪方法  

UAV Autonomy Landing Visual Tracking Algorithm

作　　者：邓红德[1] 王丽君[1] 金波 

机构地区：[1]西北工业大学无人机特种技术国家重点实验室,陕西西安710065 [2]成都飞机工业公司,四川成都610092

出　　处：《计算机测量与控制》2009年第7期1387-1389,共3页Computer Measurement &Control

基　　金：总装预研基金项目(9140A25040307HK0306)

摘　　要：视觉导航是利用图像信息进行飞机定位的,在无人机着陆过程中,为使机载传感器能始终追踪到机场跑道,提出了一种基于摄像机姿态与变焦控制的视觉跟踪方法;该方法通过调节摄像机姿态来追踪目标特征点,使目标特征点尽量位于成像平面的中心,然后根据无人机与机场跑道的距离,适时的调节焦距,以保证图像特征点完全位于图像平面内;实验结果表明,该法能很好地控制摄像机姿态角,使目标特征点位于图像中心附近,达到很好的跟踪效果;且该法操作简单,不需要增加图像处理的难度,实时响应速度快,可以满足工程需要。Vision navigation uses the image information to evaluate the aircraft position, in the process of landing, for airborne sensor can always follow up to the airport runway, presents a visual tracking algorithm which based on camera controlling. The algorithm by adjusting the camera to track the target feature makes the feature points as possible in the centre of imaging plane, and then according to the distance from the airport runway, timely adjust the focal length to ensure the image feature points in the image plane completely. Experimental results show that the law can control the camera angle well that made the target image feature points in the vicinity. And that the method is simple, don＇t need to increase the diffieuhy of image processing, and the real--time response speed is quick, can meet the project needs.

关 键 词：特征点跟踪 变焦控制 视觉导航 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]