Non-parametric Adaptive Method for Enhancing Image Resolution

An image resolution enhancing method, which is applied to the two-dimensional (2-D) frequency domain data of the original image, is proposed. The proposed method employs the narrowband filter to filter the partitioned snapshots and utilizes some weight to weight the filter outputs to obtain the amplitude estimate at special image pixel. Due to allowing the sample covariance matrix singular, the larger tap filter, which may be higher resolution, can be employed. Again, with considering the correlation among the snapshots, a new weight is proposed for weighting the filtered data. Because the proposed method fully utilizes the characteristic of the data set, it is superior to the similar method. The simulation and experimental results demonstrate the performance of the new method.