Abstract
In this thesis, an edge-sensitive demosaicking (ESD) algorithm and an effective demosaicking and zooming (EDZ) algorithm are proposed. To improve quality of reconstructed full-color images from color filter array (CFA) image, the ESD algorithm analyzes the neighboring samples around the missing green pixel to determine suitable samples (i.e. direction) to interpolate the missing green pixel. Following the interpolation operations of the green plane, the ESD algorithm makes use of the high correlation between R, G, and B planes to produce the red and green/blue and green color different planes to reconstruct the red and blue planes. Obviously, the color different plane not only looks smooth in a local area but also exploits the advantage of the green plane to reconstruct the red/blue plane more accurately. Hence, the EDZ algorithm exploits the red and green/blue and green color difference planes, and develops expressions to join the horizontal, vertical, and diagonal neighboring information to improve the accuracy of the estimated value, and then the zoomed full-color images are more pleasing to human visual. The experimental results demonstrate that the proposed algorithms have impressive performances. Keywords: Color filter array, demosaicking, interpolation, zooming