Research On Image Clearness Technology Under Complex Weather

In view of the complicated weather conditions such as fog and haze,the attenuation of light involved in imaging is caused by the scattering of tiny particles in the air which resulting in the contrast of the acquired image to decrease,and the quality of the image to decreases.It not only affects the visual effect of the image seriously,but also increases the difficulty of extracting the feature information.Serious more likely to make some outdoor monitoring system doesn't run properly.Therefore,it is practical significance to study the technology of image sharpening under the complex weather condition.It is the two kinds of method that is focused on at present,Image defogging method based on non-physical model(the image enhancement method)and image dehazing method based on physical model(the image restoration method).Image enhancement is the processing method which highlights some of the information in a specific image based on the need,and weaken or remove some unwanted information,easy-to-use and strong real-time property,but it is likely to cause the color distortion;However,the image restoration method is study to the physical model of haze imaging,clear image is obtained by assuming prior information and solving model parameters,the texture details of the fogging image are well restored,and the fog free and clear image is obtained.The texture details of the fog image are well restored,but the algorithm complexity is high and the brightness and hue of the image are low.Therefore,in the process of sharpening the haze image,it is not only to take into account the problem of image enhancement degree and detail information restoration,but also pay attention to color distortion problem after dehaze image.Based on the knowledge above,in the light of the problem of existing in the two kinds of defogging methods.We put forward an improved method,during my postgraduate period,the following work was completed:Through in-depth analysis of the shortcomings of two kinds of defogging models,we propose an improved research on single image haze removal algorithm based on parameter optimization search of linear model.First,starting from the mechanism of the model,according to the linear nature of the two kinds of methods and considering the noise influence in the process of image,a unified description form of the linear model is given.Then,in solving the problem of model equation,using the prior of estimating the parameters,the optimization ability of the genetic algorithm itself is fully utilized.The restoration of haze image is transformed into the optimal parameter estimation for the original,and the optimal search of the model parameters is realized.Finally,the optimal parameters are brought into the linear model equation and the best dehazing image is restored.The experimental results show that the edge of image are clearer,the scene is more natural,the color information is more abundant.By comparing the effect of existing haze removal methods,and evaluating the image quality.We propose a self-adaption single image dehaze method based on clarity-evaluationfunction of image,aiming at the poor universality of the image sharpening effect of different types of haze images.First,the haze image is classified by calculating the clarity-evaluation-function of the input image.Then,the adaptive dehaze algorithm is selected and carried out in turn.In order to make the hue and contrast of the dehazed image better,we introduce the coding decision method to the result of the image quality evaluation.Finally,output the best results of multiple methods to achieve the overall optimization of haze removal system.Experimental results show that the algorithm has high contrast and good visual effect,the algorithm is more universality.This algorithm can achieve haze image clarity.