Research On Low Bit Rate Compression And High Robustness Transmission Of Underwater Images

Underwater image communication plays an important role in tasks such as underwater rescue,seabed geomorphological exploration,and marine safety monitoring.However,unlike the wireless channel in the air,because of the limited underwater acoustic channel,it can only provide low bit rate data transmission.Besides,the serious multipath effect of the underwater acoustic channel will cause serious inter-symbol crosstalk,and the ocean noise interference will cause serious errors at the receiving end.In summary,there are two main difficulties in establishing an efficient and reliable underwater image communication system.One is to compress the image with the highest possible compression rate under the premise of ensuring the quality,so that it can be transmitted in an underwater acoustic channel with a limited transmission rate.The second is to establish a reliable transmission protection strategy to improve transmission robustness,so that the receiving end can reconstruct images with higher quality.This article focuses on underwater image communication.Focusing on the above two difficulties,we design an underwater image compression algorithm based on perceptual quality by combining underwater image characteristics and human visual characteristics.And combine with the transmission strategy of unequal error protection to ensure that the receiving end can reconstruct the image with high quality.In addition,we also designed a highly robust underwater image compression algorithm based on a autoencoders.The encoder itself has excellent error resistance performance,which is better suited for underwater acoustic channels with high bit error rate.The work content and innovations of this article are as follows:(1)Investigate the progress of related research at home and abroad,analyze and discuss the performance of some existing algorithms under the characteristics of underwater acoustic channels,and explain some of the shortcomings of the existing algorithms and problems to be solved.(2)Aiming at the inherent characteristics of underwater images such as low contrast,single color information and blurred details,combined with the concealment characteristics of human vision,this paper proposes an underwater image compression algorithm based on perceptual quality.By adding saliency detection and hybrid wavelets and DFBs transformation,more redundancy is removed on the premise of ensuring image quality,so as to adapt to low-bit-rate underwater acoustic channel transmission.(3)Aiming at the generally poor error resistance performance of existing compression algorithms,this paper proposes an underwater image compression algorithm based on autoencoders.Compared with other compression algorithms based on wavelet and deep learning methods,this algorithm has better error resistance and robustness.It can reconstruct better quality at the receiving end without any channel coding protection.This algorithm is suitable for underwater acoustic channels with high bit error rate.(4)Aiming at the proposed underwater image compression algorithm based on perceptual quality,a digital transmission mechanism based on unequal error protection and a pseudo-analog transmission mechanism based on unequal error protection are designed to protect the image transmission.Both of these two algorithms improve the robustness of underwater acoustic channel transmission,so that the receiving end can reconstruct images with high quality.