Real-time Image Stabilization Method Based on Low Altitude Fixed Wing UAV

Abstract

The existing electronic image stabilization techniques are mainly based on feature point matching, which is difficult to solve the problem that the low altitude fixed wing UAV Using strapdown camera platform collects fewer image feature points. To solve this problem, a real-time image stabilization method based on attitude angle information is proposed in this paper. Firstly, the attitude angle information collected by IMU and the image information collected by visual sensors are aligned in real time. Then, the attitude angles of pitch, roll and yaw are filtered differently, and the corresponding basic matrix is calculated according to the processed angular information. Finally, affine transformation is used to stabilize the original image sequence, and a smooth image sequence is obtained. The experimental results show that the peak-signal-to-noise ratio of the image sequence after image stabilization is increased by more than 5%. The algorithm is suitable for low altitude fixed wing UAV platform. For images with less feature point information, it can also eliminate jitters. At the same time, it has strong robustness to large maneuvers of aircraft.