摘要
提出一种统计插值算法 ,用于半象元错位双幅图像的复原计算。分析了半象元错位双幅图像合成后的数字特征、复原高分辨率图像的合理性及插值位置的规律 ,给出相应数据结构的矩阵表达式。在考虑实际噪声的条件下 ,由贝也斯公式出发 ,导出估计插入值的统计计算公式。该公式中用到噪声和复原图像的协方差矩阵 ,后者采用Marcov模型。实验和计算机仿真的复原图像表明 。
A statistical interpolation algorithm used for restoring bi-images of half pixel offset is introduced.Analyzing the digital trait of two images fusion of half pixel offset and the orderliness of the interpolation position,the matrix formulae are given.Meanwhile the rationality of the high-resolution image restoration is discussed.A statistical expression for estimating the interpolatory data is derived from Beyes rule.The autocovariance of noise and autocovariance of restoration data are employed in this expression.The Marcov model is used for substituting the unknown autocovariance of restoration data.The results of experiment and computer simulations show this statistical interpolation algorithm is feasible and efficient.
出处
《激光杂志》
CAS
CSCD
北大核心
2003年第4期25-28,共4页
Laser Journal
基金
省重点实验室资助项目 (KJSO10 0 3 )
