基于视觉信息的无人机自主着陆过程姿态角简化计算方法

Research on Estimation of Attitude of Vision-based Autonomous Navigation to the Landing of the UAV

研究了一种基于地平面算法的无人机着陆过程中姿态角估算的改进方法。通过地面坐标系到摄像机坐标系的转换矩阵、机体坐标系到摄像机坐标系的变换矩阵、机体坐标系到地面坐标系的变换矩阵求解,确定无人机的姿态信息;利用无人机着陆过程中跑道的图像地平线和跑道数据等特征提取,在获取摄像机姿态角的基础上,计算无人机的滚转角和俯仰角,用跑道特征来估算无人机偏航角;通过矩阵变换解算出无人机的姿态角,并通过仿真进行了实验验证。该方法对摄像机的安装位置没有特殊的要求,解除了传统方法对摄像机安装位置的限制,简化了无人机姿态信息的求解方法。

Studies were made for improving the estimation of attitude angles during the landing of the UAV based on a new horizon detection algorithm. We used the transfer matrixes from ground coordinate system to the camera coordinate system, body coordinate system to the camera coordinate system and body coordinate system to the ground coordinates to determine the UAV attitude. After extracting the features of horizon and the runway from the pictures and analyzing the relationships between the horizon and runway characteristics and UAV attitude angle, we estimated the UAV roll angle and pitch angle and used the characteristics of the runway to estimate the UAV yaw angle on base of acquiring the attitude an-gle of the camera. Matrix transformation is used to calculate the attitude angle of the UAV and stimulation was carried out to verify the results. It is estimated that there is no special requirements on the installation position of the camera. As a result, it is feasible to remove restrictions on the installation location of the camera and simplify the method to estimate the UAV attitude information.