GP-grid image interpolation and denoising for division of focal plane sensors

摘要

Image interpolation and denoising are important techniques in image processing. Recently, there has been a growing interest in the use of Gaussian processes (GP) regression for interpolation and denoising of image data. However, exact GP regression suffers from O (N~3) runtime for data size N, making it intractable for image data. Our GP-grid algorithm reduces the runtime complexity of GP from O (N~3) to O (N~(3/2)). We provide comprehensive mathematical model as well as experimental results of the GP interpolation performance for division of focal plane polarimeter. The GP interpolation method outperforms the commonly used bilinear interpolation method for polarimeters.