Abstract: As most enhancement methods do not consider the simultaneous relationship among images, light intensity, and shooting times, this paper proposes an enhancement algorithm for all-weather intersection traffic images based on adaptive brightness baseline drift (ABBD) according to illumination variations at different times. The algorithm establishes a brightness benchmark curve according to the changing illumination at different times, and then an adaptive brightness benchmark model is set up based on the brightness benchmark curve and the real-time brightness feedback. The proposed algorithm uses the brightness benchmark model to enhance the brightness of the image adaptively. Experimental results show the method's effectiveness for image enhancement in all-weather and different light conditions, as well as its robustness under different weather conditions.