融合地面多传感器信息引导无人机着陆

针对无人机自主着陆过程中卫星导航系统易被干扰的问题，提出了一种基于地基多传感器融合的无人机自主着陆引导方法。采用主动式激光照射的方式，获取机载反射棱镜的近红外成像，在红外图像中对无人机目标进行识别；通过坐标转换将识别结果映射到可见光图像中，在可见光图像中选择的感兴趣区域进行可见光目标识别，从而在降低计算量的基础上获得更加精确的无人机相对角度信息；利用距离测量信息和引导系统角度信息可以获得精确的无人机相对位置。无人机着陆引导试验结果表明，该方法能够提供精确的无人机位置信息，能有效适应于复杂背景下的无人机自主着陆引导。
A ground-based multisensory fusion approach was proposed for the guidance of UAVs (unmanned aerial vehicles) landing in GNSS (global navigation satellite system) denied environments. Firstly, an active laser transmitter was used to irradiate the UAV. The light spot in the obtained infrared images reflected by the airborne reflection prism was recognized as the UAV. Then, a region of interest was defined in the visible-light image by coordinate transformation according to the result of the infrared image. To reduce the computation complexity, the UAV was detected and located in the region of interest. Finally, the relative position of the UAV can be obtained by combining the distance measurement and the angle of the pan-tilt unit. Results of flight experiments demonstrate that the proposed approach can offer the precise positional information of UAV and can effectively adapt the complex background of UAV autonomous landing.