Kryvenko S. Methods for compression and filtering of remote sensing data with providing high visual quality

The object of research - one-and multi-channel images generated by a number of aerospace imaging systems. For the first time, a method for automatic image compression with losses without introducing visually noticeable distortions which is applicable for single and multi-channel RS data. For the first time. it is shown that lossy compression of images corrupted by intensive noise, can provide improvement of visual quality of compressed images and the method that implements compression in the neighborhood of OOP. HVS-metrics are improved allowing better accounting human attention to locally-active fragments in image analysis. Image filtering methods based on DCT are modified allowing to suppress non-stationary noise with a limited a priori knowledge about their characteristics.