散焦和运动模糊复合模型及参数估计

Hybrid model and parameter estimation combining defocus and motion blur

为了解决散焦和运动模糊同时存在的图像复原问题,建立了一种散焦和运动模糊复合模型,提出了一种基于自相关的参数估计方法。在分析图像成像过程的基础上,对标准散焦模型和运动模糊模型进行卷积,估计复合模型的分布范围,根据光的散射特性和能量守恒调整复合模糊模型的幅度。对复合模型按两方向分别求导,根据导数矩阵自相关的曲线特性估计模糊参数。实验结果表明,参数估计方法具有较高的估计精度,解决了混叠现象造成的干扰问题,复合模型则能够取得较好的复原效果。

To restore images degraded by defocus and motion blur,a hybrid model combining defocus and motion blur is established,and a parameter estimation method based on autocorrelation is proposed.Based on the analyzing of imaging process,the convolution of the classic defocus model and motion blur model is performed to evaluate the hybrid model distribution,and the hybrid model range is adjusted according to the light scattering and energy conservation.The blur parameters are estimated using the autocorrelation curves of the derivate matrix that is calculated from the hybrid model in two directions.The results show that the proposed estimation method his more precise,which overcomes the interference by overlapping,and that the restoration by the hybrid model has a jood effect.