Abstract: In order to remove fog for single image fleetly, a defogging method based on dimension reduction filter is proposed in this paper. It uses boundary constraint to achieve transmission map, and then optimize the transmission map with dimension reduction filter. The complex computation is reduced considerably. Distortion of the image in the sky area is compensated by adaptively adjusting the atmospheric optical value and lower limit of transmission map. Then the visual effect of optimized image is improved by applying an adaptive Zadeh-X transformation. Experimental results on a variety of haze images demonstrate real-time performance and high-quality of the proposed algorithm, and it is applicable to foggy scenarios despite illumination is too low or too high and the color information is not rich enough