In the battlefield, fog and haze lead to image degradationand, which affects the target recognition and command decision. Therefore a battlefield image defogging algorithm, which combines bilateral decomposition and $L_1$ dark channel, is proposed. The bilateral filterer is used to decomposed the original image into low frequency component with fog information and high frequency component with noise image. For low frequency image, the sky area is split to obtain the global atmospheric light, and $L_1$ regularization is built according to the context constraint equation for accurate transmittance to improve the dark channel prior model for the fog. For high-frequency images, the guide filter is used to restrain noise and enhance detail, and then the final image is obtained by image fusion and bi-gamma function correction, so as to realize the decoupling of image fog removal and noise removal. The proposed algorithm is compared with the existing algorithms, and applied to the defogging of the image in the battlefield environment, which is analyzed using the objective evaluation factor. Experimental results show that the proposed algorithm can effectively remove haze, suppress noise, enhance visual effects and is better than the existing algorithms in objective evaluation factors.