无人飞行器自主着舰实时场景的仿真实现

为给垂直起降无人飞行器视觉自主着舰算法研究提供经济可行的解决方案,实现了无人机自主着舰场景的软件仿真以提供算法研究中所需的合格场景图片.基于机器视觉理论,通过分析无人飞行器机载摄像机在自主着舰时考虑了透镜畸变参数的针孔成像模型,给出了无人飞行器着舰靶标的成像公式,并介绍了软件仿真中畸变模型实现时出现的问题及解决办法.为了使仿真图像更逼真,综合考虑了无人飞行器着舰实时场景中所存在的主要干扰,如光照不匀,运动拖影及电气噪声和大气流动和海面悬浮颗粒造成的白噪声等,在C+ +和Windows环境下实现了无人飞行器自主着舰的仿真图像和数据的实时生成.最后展示了部分仿真结果.

Software imitation for unmanned aerial vehicle landing scenes

To offer a cost-effective solution to the research of vision-guide algorithm for UAV(unmanned aerial vehicle) with vertical takeoff and landing capability, the imaging principle and imitation technology of UAV landing scenes based on the theory of machine vision was employed. The pinhole camera model including lens distortion parameters was established. Moreover, a method of adjacent region search was introduced to decrease the singularities in the imitation images caused by the lens distortion. To get more veracious imitation of landing scenes, some main interferential factors of the unmanned vehicle landing, such as the blurring induced by the light amplitude non-uniformity, image motion, as well as the noise from electronic and photometric source, were also considered. The software for real time generation of scene images was developed by C and windows operating system. The results of simulation including images and data are presented at the end.