Colour demosaicking method using adaptive cubic convolution interpolation with sequential averaging

Abstract

Colour demosaicking refers to the process used to estimate missing colour components at each pixel position in a single-sensor imaging device equipped with a colour filter array. The proposed method starts the demosaicking process by estimating the luminance channel, which utilises cubic convolution interpolation along the direction of the smallest gradient magnitude at each chrominance pixel position. Reconstruction of chrominance channels comprises spectral correlation-based average of neighbouring chrominance pixels and sequential filtering on the reconstructed chrominance channels. The sequential filtering can be regarded as a postprocessing stage and effectively smears out colour artefacts. Experimental results show the effectiveness of this approach.