飞行移动平台成像匹配的电子稳像技术研究

Speeded-up Robust Feature Matching for Video Stabilization Technology

针对移动飞行平台飞行中拍摄视频序列存在平移,旋转,缩放等随机运动,并易受光照变化,噪声影响,难以实现目标的跟踪匹配的问题,提出一种快速有效的基于改进加速鲁棒性特征(SURF)匹配的电子稳像算法。上述方法首先通过SURF算法提取图像中特征点及描述;然后提出一种双向匹配策略获取特征匹配点对;再采用改进的随机采样一致性算法(RANSAC)进一步剔除误匹配点,最终利用计算出的仿射运动参数对抖动帧进行补偿,实现视频序列的稳像处理。实验结果表明:该方法能有效补偿原始视频序列存在的复杂随机抖动,鲁棒性好,稳像速度<30ms,输出视频序列峰值信噪比(PSNR)平均提高5d B,为提高移动视频质量提供了参考。

In the paper,a fast and efficient video stabilization method was presented based on speeded-up robust features（ SURF） matching. Firstly,the SURF algorithm was used to extract the fearture points and obtain their descriptors in each frame. Then the feature matching points were acquired with two-way matching strategy. After that,the improved random sampling consensus algorithm（ RANSAC） was used to further eliminate false match points; and the calculated affine motion parameters was used to compensate for the effect of random jitter to achieve image stabilization processing of the video sequence. The experimental results show that this method can effectively compensate the original video sequence of complex random jitter,the robustness is strong,the speed of video stabilization is under 30 ms,and the output video sequence peak signal-to-noise ratio（ PSNR） average increase 5 db.