面向图像局部重建的梯度分析

Gradient Analysis for Image Local Reconstruction

在基于场方向导数的图像局部区域重建过程中,梯度计算对图像边缘的重建效果有很大影响。分析选择单向计算梯度分量的必要性,提出梯度计算的就近原则,对参与梯度分量单向计算的相邻点设置优先级,将缺损点与已知点之间的边缘特征反映在已知点的梯度中,进而影响重建过程,从而实现对图像中边缘特征的重建。实验结果表明,采用该梯度计算模式能有效改善对图像边缘特征的重建效果。

In the reconstructing image local region based on directional derivative, gradient calculation plays an important role in improving the effect of the edges ’ reconstruction. Gradient component unidirection calculation is chosen,and adjacency principle is proposed to choose an adjacent point between the two to calculate the gradient component. The edges between the lost points and the known can be reflected by the known gradient in time,and it can be reconstructed accurately. Experimental results show that using the gradient calculation model can effectively improve the reconstruction effect of image edge character.