A high-resolution radiospectrograph image reconstruction method

A new type of image reconstruction method is introduced. The image reconstruction is based on an adaptive regression type of destriping method using a regression model prediction for replacing a corrupted image row representing a frequency channel or isolated pixels. The method uses available information from the pixels next to the corrupted ones thanks to the spatial correlation of the radiospectrograph data. The method suggested can be successfully applied also to other image acquisition and processing applications, i.e. remote sensing, pattern recognition, meteorology, medicine, etc.