基于视觉系统技术的无人机自主定点降落与实例验证

Autonomous fixed-point landing of UAV based on vision system technology and case verification

为解决无人机定点降落效果不好问题。提出视觉特征融合下的电力巡检无人机单兵自主精准定点降落算法。依据无人机与地标点的空间位置关系,对着陆区域位置进行估计,并通过计算数字图像的空间矩,对降落区域轮廓进行多级筛选,以检测着陆区域轮廓,并利用线性方程求取无人机与降落区域距离的残余误差,以此为依据,采用视觉特征融合算法获取无人机降落过程中的估计量,进而计算得出无人机单兵相对于降落标志的最优偏差角度,由此实现电力巡检无人机单兵自主精准定点降落。应用结果表明,所提方法对于地面降落区域的降落点识别效果较好,定点降落性能更佳。

In order to solve the problem of poor fixed-point landing effect of UAVs,an autonomous and precise fixed-point landing algorithm for individual soldiers of power inspection UAV based on visual feature fusion was proposed in this paper.Based on the spatial position relationship between the drone and the landmark,the landing area position was estimated.By calculating the spatial moment of the digital image,multi-level filtering was performed on the landing area contour to detect the landing area contour.Linear equations were used to calculate the residual error of the distance between the drone and the landing area.Based on this,a visual feature fusion algorithm was used to obtain the estimation during the drone landing process,Furthermore,the optimal deviation angle of the drone individual relative to the landing sign was calculated,thereby achieving autonomous and precise fixed-point landing of the electric power inspection drone individual.The application results of the example showed that the proposed method had better recognition performance for landing points in ground landing areas and better performance for fixed-point landing.