An Image Interpolation Method with Edge-Preserving

Abstract

In image processing, the low resolution problems of imaging system are solved by interpolation techniques. In this paper, we propose an edge-preserving image interpolation method to interpolate the image filmed by linear array camera. For a 1D signal, many signal segments are extracted by a processing window sliding along the signal sequence. For one missing data sample, there are a lot signals segments can be used to recover it. These signal segments are approximated by polynomials, in order to preserve the edges, the one with the minimum fitting error is used to estimate the missing data sample. For a 2D image data, a lot of 1D signals are extracted along directions specified firstly, and then interpolation results are obtained by applying above 1D edge-preserving interpolation method on all these 1D signals. The one with the minimum fitting error is selected as the final output. Experimental treatment results indicate that the proposed method is edge-preserving, and can be used to reconstruct the image filmed by linear array camera effectively.