The Research Of Single Image Clearness Algorithm In Haze-fog Days

In recent years, haze has become a new type of environmental disaster. Due to the scatter of atmospheric particles, the images obtained by image acquisition device have low contrast, image blurring, and details information loss, etc. Therefore, it has important theoretical guidance and practical significance to study the characteristics of fog particles and find an effective and fast defog method to sharpen images. This paper, based on the research of fog particles and image enhancement methods, discusses how to clear color image in the weather of fog and haze.Firstly, from the point of characteristics of haze particles, this paper mainly studies the component of atmospheric particles, the reasons for the formation of haze weather conditions and its impact on image acquisition device, etc. The extinction characteristics of single and mixture of particles with different size and refractive index are analyzed. The calculation shows that the extinction efficiency factor of single PM10 particle changes with the different particle size and refractive index. And the extinction efficiency factor of a lot of particles increases rapidly with the change of particle sizes firstly. Then it will decreases and tends to a constant. The factor is also associated with the change of the width of particle size distribution and refractive index, usually not monotonous. For external mixture of particles, the extinction curve also relates to the mixing ration for the mixing particles.Secondly, this paper analyses the enhancement algorithm of histogram equalization which is convenient and direct. At the beginning, global histogram equalization algorithm is adopted to the haze image processing. Aiming at the shortcoming of this algorithm like the loss of image details and image blurring, partially overlapped histogram equalization and linear filtering method are employed to overcome them. To a certain extent, it can eliminate the block effect by partially overlapped histogram equalization, greatly enhance the details of the image, and improve the overall clarity.At last, the de-haze algorithm based on Retinex algorithm is researched in depth. The algorithms like SSR, MSR and MSRCR are simulated and analyzed in this paper. Combined with the adaptive linear stretching method, image contrast is further improved. The results show that the algorithm can improve the overall image contrast, maintain the image colors better and achieve ideal clear effects.