In order to reduce the redundancy of image data caused by additive noise introduced by aging image acquisition equipment, the effect of additive noise on image and video is studied, and the BM3D (block matching 3D) algorithm to process noisy static images is proposed in this paper. It is verified that the image information entropy and data volume can be reduced by reducing noise. Then the noise reduction-coding scheme is proposed. In this scheme, the noisy video is firstly denoised by VBM4D algorithm, and then encoded by H.264. After the scheme processing, the image information entropy is reduced by 16%, the data volume is reduced by 38%, and the encoded code stream is reduced by 59.5% on the premise that the video quality is not significantly reduced. The data show that the encoding compression rate of H.264 can be significantly improved and the transmission of video in small bandwidth communication systems is promoted by this method under the premise of guaranteeing video quality. Compared with H.264 and neural networks, this scheme has low complexity and can process video relatively in real time.