单幅彩色运动模糊图像的参数估计

Parameters Estimation from a Single Motion Blurred Color Image

提出一种基于单幅彩色模糊图像的运动参数估计方法。首先分析运动模糊图像的傅里叶频谱,揭示图像蕴含运动估计所需要的信息;其次对彩色图像采用分通道估计,避免图像转化带来的信息损失。针对参数估计中误差叠加,对运动方向和尺度分开估计。根据Gabor滤波器对模糊图像频谱图的响应关系得到运动角度;用倒谱法分离出模糊图像中的运动信息求得运动尺度。实验结果表明,该方法可以更精确地从单幅图像中估计运动参数。

A method of motion parameters estimation based on a single color motion-blurred image was proposed. Firstly, the Fou- rier transform of the motion blur image was analyzed, thus the motion information for motion estimation in blurred image was obtained. Secondly, the three channels were treated separately for color image motion estimation to avoid the loss of information by image conver- sion. Aiming at the error of parameters getting stacked, the length and orientation of the motion were estimated respectively. The orien- tation of motion was estimated from the response of the Gabor filters which related with the spectrum of the blurred image. Additionally, the motion extent was obtained form the motion information of the blurred image by using cepstrum method. The experimental results show that this method is more precise for the motion parameters estimation based on one single blurred image.