Outlier Removal for Super-Resolution Problem Using QR-Decomposition.

摘要: Super-resolution algorithms reconstruct a high resolution image from set of low images scene. Recent advances in Super-Resolution techniques show trends towards robustness against out- liers. Even under normal lighting condition, may have motion outliers due to independent moving objects. In this article new method for dealing objects is proposed, which those frames containing are identified using QR-Decomposition, known Linear Algebra. This detection procedure comes between the registration and reconstruction steps super-resolution techniques. The detected as outlier will be dropped hence step only works on other frames. simulation results better performance proposed algorithm compare with some methods.