Research On Fog-degraded Digital Images Processing

Because of the atmosphere scattering phenomenon , weather degraded images captured in fog suffer from low visibility. This significantly affects the traffic systems and outdoor visual systems. Hence , it is an important significance for fog-degraded images to processing. At present,researches on the enhancement of mist-degraded images are still under discussion. A few researches are concentrating on the analysis of the complex physical model. Due to the complexity of modeling situations,many of these methods are not convenient.This paper takes the enhancement procedure as an image-processing problem. Aiming at the images that received in the fog with visibility 500m - 1,000 meters, processing methods are researched and quality of the images is assessed. Our work includes:1. The method of edge extraction is used to retain related edge information for the mist-degraded images. Theories and algorithms including binary wavelet edge detection,Edge fusion theory,the image edge detection algorithm based on multi-scale edge fusion are deeply analyzed in detail.2. According to the characteristics of noise of the mist-degraded images, the wavelet threshold de-noising methods are analyzed. The application of wavelet transform in image de-noise is studied by threshold operation in wavelet domain of image. And the method that can be used to estimate the threshold value was given.3. Aiming at the massive phenomenon and loss of pixels of the images, 3-spline interpolation is used to increase pixel for wavelet threshold de-noising images and to smooth images. This method avoids the massive phenomenon that the traditional method creates. 4. The application PSNR and information entropy is for the quantitative analysis, comparing with the results of histogram equalization.Theoretical analysis and experimental results demonstrate that method sued in this paper are effective for contrast enhancement of mist-degraded images,and can be applied to the practical situations.