针对制导弹药视觉景象匹配导航中由于飞行姿态引起的定位误差,提出了一种适用于高动态、倾斜视角情况下的惯性辅助视觉景象匹配的导航方法。该方法对现有的景象匹配导航方法进行改进,以惯性姿态辅助视觉景象匹配导航求解校正前后图像之间的单应性矩阵以及重采样尺度,进行实时图的畸变校正,避免在后续匹配算法中多次迭代盲目匹配,提高算法效率;同时,对匹配算法后得到匹配点位置信息进行定位修正,减小飞行过程中的定位偏差。实验表明:该方法能够有效减小载体倾斜在高动态环境下引起的导航定位误差。其中,图像校正算法平均误差为1.07像素,耗时为0.005 s,与传统方法相比平均误差减小69%,耗时缩短88%。惯性辅助视觉景象匹配导航系统位置误差为11.23 m(RMSE),更新频率优于10 Hz,能够有效应对中末制导的高动态、倾斜环境,为卫星拒止下的制导弹药提供低成本的自主导航方案。

Aiming at the positioning error caused by the flight attitude in a visual scene matching navigation of guided munitions, an inertial aided visual scene matching navigation method is proposed for high dynamic and inclined viewing angles. This method improves the existing scene matching navigation method. It uses inertial attitude to assist visual scene matching navigation to solve the homography matrix and the resampling scale between the images before and after correction, and corrects the distortion of the real-time images, which can avoid the blind matching of multiple iterations in the subsequent matching algorithm and improve the algorithm efficiency. At the same time, the position information of the matching points obtained by the matching algorithm is corrected to reduce the positioning deviation during the flight. Experimental results show that this method can effectively reduce the navigation and positioning errors caused by flight attitude in high dynamic environment. The average error of the image correction algorithm is 1.07 pixels and the time consumption is 0.005 s, which reduces the average error by 69% and the time is shortened by 88% compared with the traditional method. The position error of the whole navigation system is 11.23 m (RMSE), and the update frequency is better than 10 Hz, which can effectively cope with the high dynamic and inclined environment of mid-terminal guidance. It provides a low-cost autonomous navigation solution for guided munitions under satellites denial conditions