The Study Of Remote Sensing Image Enhancement Based On NSCT

Remote sensing image is a special kind of digital image for feature of ground objectsrecording. It is now being widely used in various fields. In the reception and transmission ofremote sensing images, the quality of images declined inevitably due to various internal orexternal reasons. The purpose of remote sensing image enhancement is to make the enhancedimages more suitable for the interpretation of human eyes or the analysis and processing ofcomputer by adjusting the local or global features, amplifying the difference informationbetween different target objects and suppressing redundancy and interference information.The processes and effects of image enhancement are different with applications and targets.The procedure of image enhancement may be a process of distortion. The main purpose ofimage enhancement method of this study is to make the enhanced images meet theinterpretation and identification requirements of human eyes in the aid of computer.We made our study based on existing remote sensing image enhancement methods. Firstof all, the WindowShrink and BayesShrink based on NSCT is combined to enhance images,which get good adaptive de-noising effect, but can’t enhance the image contrast and edgeinformation at the same time. Based on these, we further propose an adaptive imageenhancement algorithm based on BayesShrink. Due to remote sensing images have lowcontrast and rich detail information, the HE is utilized to enhance the contrast of remotesensing image at first. The classical histogram equalization (HE) can quickly and effectivelyimprove the contrast of image globally, however, globally-based methods often do not takeinto account the relationship among neighborhoods and leads to noise amplification and poorimprovement in local information. Then the remote sensing image after HE is decomposedinto the NSCT domain with a low-frequency sub-band and several high-frequency sub-bands.Cause noise and edge details concentrated in the high frequency sub-band, an adaptivethreshold de-noising method is used to remove noise in the high-frequency. To furtherenhance the contrast of image locally, the fuzzy contrast enhancement is applied at the high-frequency sub-bands after de-noising. The low-frequency part stays without beingprocessed to prevent over-enhanced. At last the remote sensing image is transformed intofrequency domain and a Laplacian filter is utilized to enhance the edges.At last, in order to test the performance of the proposed method and their comparedmethods, results of experiments were analyzed from both objective and subjective. Theexperimental results show that the proposed method performs better than other methods oncontrast enhancement, noise suppression and detail direction preservation.