基于增量维纳滤波的图像插值算法辨识

Image Interpolation Algorithm Identification Based on Incremental Wiener Filtering

针对数码相机中颜色滤波阵列和插值算法的辨识问题,提出基于增量维纳滤波的联合辨识方法。建立数码相机插值过程的近似模型,根据迭代增量维纳滤波的迭代误差和迭代维纳解联合估计颜色滤波阵列模式和插值系数,利用估计的插值系数实现插值算法的辨识。实验结果证明了该方法的有效性。

Aimed to identify the Color Filter Array（CFA） and interpolation algorithm in cameras, a joint identification method is presented. The approximately model of interpolation is established, under this model, the iterative error and Wiener solution from incremental Wiener filtering is exploited for joint estimation of color filtering pattern and interpolation coefficients, which is identified depending on the estimated interpolation coefficients. The performance of proposed method is analyzed from the statistical data. Experimental results show the efficiency of this method.