This paper proposes a technique for enhancing weak smoke information at the beginning of a fire by using the information between video image sequences. An improved ViBe algorithm is used to monitor the entire video and obtain the motion mask of each frame. The smoke information is amplified by utilizing the frame difference amplification algorithm that removes outliers. Moreover, improving the contrast of smoke regions through the use of an improved CLAHE algorithm. Smoke information is preserved using a merge mask, while the flip mask retains background image information. Finally, the smoke area and background area are merged to obtain the final image. Experimental results show that the algorithm can significantly improve the contrast of the smoke leaking area. As the original image background is fused, the noise is significantly suppressed.