论文部分内容阅读
为获取地标信息并计算无人机当前姿态,提出一种基于图像轮廓拟合的地标识别算法。所提方法通过施加几何约束判断地标。若图像中包含完整地标,实时的角点计算可以获取无人机的姿态信息;若只包含部分地标,则可估计出无人机的移动方位和平移量,以使地标完整出现在视野中。实验室条件下仿真表明了算法的稳定、可行。
In order to obtain the landmark information and calculate the current attitude of the UAV, a landmark recognition algorithm based on image contour fitting is proposed. The proposed method judges the landmark by imposing geometric constraints. If the image contains a complete landmark, real-time corner calculation can get the attitude information of the drone; if only part of the landmark, you can estimate the movement of the UAV and translation to make the landmark complete in the field of vision. The simulation under laboratory conditions shows that the algorithm is stable and feasible.