Research On The Method Of Enhancement Of Fog Road Image

Fog is a common natural phenomenon, but also a severe weather. Because of the water evaporates, it will also have mist, even in the sunny summer. The natural light through condensation droplets of fog will be scattered, leading to image information blurred which was caught by human eye or optical sensor, meanwhile, the contrast of the got fog image follow the scene depth to exponential decay, which had a strong impact on the security of video surveillance system. In particular, it will forms serious obstruction to vision in road transportation, and is one of the major causes of traffic accidents. Therefore, studying how to clear the degenerate road image in foggy weather conditions has important theoretical and practical significance.This paper mainly use image enhancement method to clear the fog-degraded road image, and the purpose of image enhancement is to make the enhanced image which is more suited to human observation and computer analysis. The article first analyzed the basic characteristics of fog-degraded image, such as characteristics of justification and contrast attenuation, especially expounded fog road image characteristics of histogram, gray average and variance, also proposed evaluation parameters, which provided reference for objective evaluation of fog road image.Secondly, the article applied traditional image enhancement method analyzing the degraded road image which had serious obstruction to vision, that is, fog image road of changing widely in scene depth. This paper separately carried on the enhancement processing simulation experiment to the image, which used spatial domain and frequency domain method, such as three sections linear stretching method, histogram equalization, local histogram equalization and homomorphic filtering, and analyzed various method characteristics from objective data, by comparing the experimental results, pointed out shortcomings of the traditional enhancement method such as over-enhancement, details blur and distortion.Lastly, this paper applied wavelet theory to deal with fog road image of changing widely in scene depth. Based on analyzing image feature of wavelet decomposition, the article proposed a non-linear enhancement method based of wavelet decomposition, which applied non-linear enhancement operator to deal with wavelet transform coefficients and then reconstructed image. The experimental results indicate that the method not only better overcome shortcomings of traditional enhancement method but also can well describe the image edges and details, and obtain a better effect of image clearness.