基于视频抖动的灰度投影稳像算法

Video Jitter Based Image Stabilization Algorithm

针对抖动导致车载视频图像失真的问题,将图像区域等分的思想与灰度投影算法融合,获得一种利用局部运动矢量估算全局运动矢量的方法;借助图像的全局运动矢量和滤波的思想,利用运动矢量偏差补偿抖动对图像失真的影响,获得一种稳像算法。数值实验结果表明,该稳像算法对于提高车载视频图像的稳定性具有好的效果。

This work develops an improved gray projection algorithm to estimate video image,global motion vector in terms of its local motion vectors,which integrates the idea of image region division with the basic gray projection algorithm.Simultaneously,an image stabilization algorithm is designed to solve the problem of video distortion caused by car jitter,depending on image filtering and the acquired motion vector deviation.Numerical experiments show that the proposed image stabilization algorithm can perform well with the aspect of image stability.