基于三维激光扫描的无人机自动导航技术研究

Research on automatic navigation of unmanned aerial vehicle based on 3D laser scanning

传统无人机自动导航方法依靠卫星采集的图像进行自动导航时,由于卫星采集图像范围较大,无法实现对目标区域运动目标的准确提取,容易出现导航过程耗时长和导航结果精度差的弊端,因此提出了基于三维激光扫描的无人机自动导航方法。通过三维激光扫描仪获取物体详细的空间信息,无人机自动导航技术包括无人机自动识别和运动目标提取,无人机自动识别分别采用瞬时差分法和光流法,获得静态目标和运动目标的运动区域,对超出运动区域的目标采用双背景原理进行前后景分离,得到运动目标的准确提取结果。基于该结果,采用基于模块匹配的无人机自动导航算法,对匹配的图像进行相似度计算,通过在目标识别区域进行随机搜索,实现无人机对目标的自动导航。实验结果表明,所提方法具有较高的导航效率和导航效果。

The traditional UAV automatic navigation method based on GPS technology collects image by satellite.because the satellite image acquisition range is large,UAV is unable to realize the accurate extraction for the target area of the moving target is easy to appear the problems that navigation process is time-consuming and the accuracy of navigation is low. Therefore,the UAV automatic navigation method gets the detailed spatial information of objects through 3 D laser scanner. UAV automatic navigation technology includes UAV automatic recognition and moving object extraction. UAV automatic recognition uses instantaneous difference method and optical flow method to get motion area static and moving targets. To separate the motion region beyond the target,uses the principle of double background to Seperate the foreground and background,gets the accurate extraction of moving targets. Based on this result,the automatic navigation algorithm of UAV based on module matching is applied to calculate the similarity of the matching images,and achieves the automatic navigation of UAV by randomly searching in the target recognition area. The experimental results show that the proposed method has high navigation efficiency and navigation effect.