Efficient Coding of Underwater Video using Wavelet-based Algorithm

Q.-Z. Li, W.-J. Wang, J.-X. Liu, and A.-Y.Zang (PRC)


Underwater video coding, Very low-bit-rate coding, WDR, and Discrete wavelet transform (DWT)


A very low-bit-rate coding method has been developed for use in underwater video compression to meet the limited bandwidth requirement of underwater acoustic channels. The new algorithm effectively reduces the visual redundancy of underwater images by means of wavelet-based preprocessing technique. Then both the intra-frame encoding and the inter-frame encoding are performed on the preprocessed input frames. According to the characteristics of underwater images, a Modified Wavelet Difference Reduction (MWDR) algorithm is proposed to encode the intra-frame. Motion estimation and compensation of inter-frames are performed on the preprocessed frames, thus not only greatly reducing the amount of computation for motion estimation but also improving the compression efficiency. A dual-WDR algorithm is employed to encode the residual images after motion compensation, which enables the significant coefficients to be distributed more compactly and coherently, thereby improving the coding efficiency of WDR-based algorithm. The experimental results show that while maintaining the acceptable quality of the reconstructed video, the proposed underwater video coding algorithm can provide average compression ratios from 250:1 to 500:1, basically meeting the requirement of underwater acoustic channels with transmissions rate of 16~32 Kbps.

Important Links:

Go Back