A new technique for efficient compression of remote sensing images using the Hartley transform and Matched Slope Zonal Filtering (MSZF)

A new technique is developed for zonal filtering of transformed environmental images that avoids some of the disadvantages of already existing techniques and results in better image compression. The two-dimensional discrete Hartley (DHT) and Fourier transforms (DFT), are used to investigate the performance of the new technique and compare it with that of other currently used schemes. The basic idea behind the suggested technique is that each image has an optimal slope direction choice that specifies the zone that gives the best reconstruction results. This slope is determined by the nature of the image and its statistics. So, the authors call their new technique Matched-Slope Zonal-Filtering (MSZF). It can be considered as an improved generalisation of most known zonal filtering schemes. True remote satellite sensing images are used to test the technique.