To overcome the problem of insufficient enhancement and weak naturalness preservation ability of the image in a closed cavity with uneven illumination intensity, a double exposure image fusion algorithm is proposed in this paper. Firstly, the optimal exposure is estimated by the image entropy maximization criterion. Secondly, the camera response model is used to synthesize the optimal exposure image. Thirdly, the low-light image and the optimal exposure image are fused based on the adaptive weight evaluation method. Finally, the improved BM3D (block-matching and 3D filtering) method is used to denoise the fusion image. The proposed method is validated in a closed chamber of open data set and GIS equipment. Experimental results show that the proposed method could increase the entropy of the image, reduce the noise of the image, improve the brightness of the low-light part, and protect image naturalness.