A pixel-labeling-based parameter-adaptive dehazing algorithm was proposed to address color distortion, detail blurring and low contrast commonly observed in aerial images captured under foggy conditions. The causes of image degradation were analyzed, and the dehazing task was transformed into pixel-label estimation of atmospheric light and transmission. A pixel-level label assignment model was constructed. A weighted graph optimization method was employed to adaptively adjust the labels and improve the estimation accuracy of atmospheric light value based on image brightness distribution and pixel similarity. A multi-label classification approach was applied to optimize the transmission, eliminating halo effect and edge blurring. The recovered clear images were obtained by using a foggy image restoration model. The experimental results showed that the proposed algorithm effectively suppressed noise, enhanced fine details, and improved image contrast and clarity in various aerial scenes. Significant improvements in information entropy (IE), structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR) were observed compared with existing methods, which validated its robustness and generalization capability across diverse foggy environment.