The invention discloses a single image defogging method based on edge classification weight fusion, and mainly solves the problem of non-ideal defogging effect caused by inaccurate transmittance evaluation of a foggy image in the prior art. The single image defogging method includes the steps: 1) acquiring a single foggy image by a camera; 2) classifying edges of the foggy image and acquiring deep edges of the foggy image; 3) dividing the foggy image into a deep edge area and a non-deep edge area according to the deep edges; 4) calculating correction dark primary color values of the foggy image in the deep edge area and the non-deep edge area in different modes by deep edge control; 5) calculating optimal transmittance of the foggy image according to the correction dark primary color values of the foggy image; 6) acquiring a recovery image according to the optimal transmittance of the foggy image. The acquired transmittance is more accurate, the recovery image is obvious in defogging effect, and the single image defogging method has the advantages of high contrast ratio and natural and true color tone and can be used for an outdoor imaging system.