Bayer pattern based digital zooming approach

In this paper, a new zooming algorithm suitable for single-sensor digital cameras is introduced. The proposed method is capable of zooming and enlarging Bayer data acquired by single-sensor cameras. The approach allows for operations on noise-free data at the hardware level. Complexity and cost implementation are thus greatly reduced. The proposed method utilizes a color difference model and an adaptive edge sensing mechanism capable of tracking the underlying structural content of the Bayer data. Simulation studies indicate that the new method yields excellent performance, in terms of both subjective and objective image quality measure.