Wireless video transmission over MIMO-OFDM using background modeling for video surveillance applications

In this paper image content analysis and adaptive background modeling is used to develop a video surveillance system that is suitable for wireless video transmission networks. The proposed system can localize any important information from the video data, i.e. location of a person, and then transmit the important information to the receiver side at the same time use background modeling to reduce the amount of video data that needs to be transmitted. In our experimental system, Motion JPEG video compression and 2×2 MIMO-OFDM wireless systems were used as the base system. The wireless channel environment is simulated using indoor and outdoor channel models. Simulations show that the system can reduce the amount of video data of up to 87.55% with average PSNR of 25.12 dB using SD resolution images; this is comparable to the image quality when background modeling is not used. Simulations show that the proposed scheme achieves slightly lower PSNR performance but with significant reduction in the amount of transmitted video data.