Quantitative Image Restoration

摘要

Even with the most extensive precautions and careful planning, space based imagers will inevitablyexperience problems resulting in partial data corruption and possible loss. Such a loss occurs, forexample, when individual image detectors are damaged. For a scanning imager this results in missinglines in the image. Images with missing lines can wreak havoc since algorithms not typically designedto handle missing pixels. Currently the metadata stores the locations of missing data, and naivespatial interpolation is used to fill it in.Naive interpolation methods can create image artifacts and even statistically or physically im-plausible image values. We present a general method, which uses non-linear statistical regressionto estimate the values of the missing data in a principled manner. A statistically based estimateis desirable because it will preserve the statistical structure of the uncorrupted data and avoid theartifacts of naive interpolation. It also means that the restored images are suitable as input forhigher-level statistical products.Previous methods replaced the missing values with those of a single closely related band, byapplying a function or lookup table. We propose to use the redundant information in multiplebands to restore the lost information. The estimator we present in this paper uses values in aneighborhood of the pixel to be estimated, and propose a value based on training data from theuncorrupted pixels. Since we use the spatial variations in other channels, we avoid the blurringinherent spatial interpolation, which have implicit smoothness priors.