Temporal Masking with Luma Adjusted Interframe Coding for Underwater Exploration Using Acoustic Channel

Abstract

Transmission of video information from a submerged sensory source may lead to numerous applications such as surveillance, deep sea exploration, monitoring and repairing of underwater oil pipes etc. Such application requires a real time video transmission. It is obviously a challenging task to achieve efficient underwater video transmission because of the limited channel bandwidth and water is not a uniform medium due to which the transmitted signal is more error prone. Due to the limited spectrum, transmitted information must be compressed in a way that its quality does not deteriorate and at the same time, it must achieve low bit rate. In this regard, we propose an algorithm of temporal masking with Luma adjustment to reduce the video information for efficient transmission and reduce the need of over compression. Proposed algorithm is applied on the moving picture expert group (MPEG) video standard. A 2D-least mean squared adaptive filter (LMSAF) is embedded at the receiver end in order to reduce the effect of aquatic noises and further compensate the overall quality. Motion in the video sequence is exploited to extract motion vectors for temporal masking and also to achieve desired bit rate for acoustic applications. Structural Similarity Index Matrix (SSIM) acquired by implementing the proposed algorithm is 0.908 while the standard H.264/AVC was 0.925, but the compression ratio was reduced to 5.03 in comparison to 7.33, which clearly shows the significance of the proposed algorithm in reducing the information size.