Edge Detection on the Bayer Pattern

This paper describes a novel investigation of edge detection on Bayer-patterned images. Inspired by the color filter array (CFA) interpolation method, we design kernels to perform Gaussian-like smoothing and Laplacian-like edge detection on a Bayer-patterned image. Benefits of using our proposed kernels to process such image source include the computation saving during the interpolation and/or the color space transformation to a full color or grayscale image, and also lower memory requirement on the same time