基于感知质量驱动的水声图像传输与优化技术研究

It has always been the major concerning in the underwater acoustic image engineering research to find out the most suitable balance between the validity and the reliability in the underwater image communication. Driven by what kinds of metric for the design and optimization of the system would influence the performance and efficiency and should be questioned very seriously. In this project, transmission and optimization method for underwater acoustic image based on perceptual quality driven was proposed, which introduced the image quality perceptual technology to the underwater image’s transmission and optimization. By taking the characteristic of the human visual system that would be patience to some image distortion into account, it could be able to allow some kinds of the image transmission error, which maybe not easily to be perceived by human eyes. Using the technique about the perceptual quality driven to guide the research on image compressing, transmission, and error repairing, the overall performance of the underwater acoustic image communication would be enhanced and achieved the balance between the validity and reliability. The research content was included as this: 1) Application about the image quality aware in the transmission and optimization for underwater acoustic image. 2) Research on the compression method for underwater acoustic image based on perceptual quality driven. 3) Research on the Transmission method for underwater acoustic image based on perceptual quality driven. Through the research of this project, new idea and method would be promised for the transmission and optimization of the underwater acoustic image, and it would also be able to introduce the technique not only for the underwater video, voice and so on, but also for other kinds of multimedia communication application.

在水声图像通信的有效性与可靠性之间找到最优平衡点一直都是水声图像工程研究的重大关切点。以什么为评价依据来驱动优化设计决定了系统的性能与效率，是需要认真探究的核心问题。本项目提出基于感知质量驱动的水声图像传输与优化方法，将图像质量感知技术应用到水声图像的传输与优化中，利用人眼视觉对图像失真存在一定包容性的特点，选择性地容许不易被视觉感知的图像传输误码的存在。基于感知质量驱动技术，指导图像的压缩、传输与修复，能实现有效性与可靠性的优化平衡，提高水声图像通信的整体性能。研究内容包括：1）图像质量感知在水声图像传输优化中的应用；2）基于感知质量驱动的水声图像压缩方法研究；3）基于感知质量驱动的水声图像传输方法研究。本课题的研究将为水声图像的传输与优化提供新思路和新方法，并为水下视频、语音等多媒体通信提供必要的技术支撑。

“基于感知质量驱动的水声图像传输与优化技术研究”课题组围绕项目拟解决的核心问题，全面搜集整理相关研究文献，密切跟踪本领域的技术发展动态，系统梳理研究关键理论方法，认真执行项目计划书的内容，较好地完成了项目研究计划和工作任务。研究工作和成果达到了预期目标。主要工作有：.(1)针对“图像质量感知在水声图像传输优化中的应用”，系统整理了建立了水下图像数据库，并结合水下图像数据库的特点，提出并实现了F.R， R.R和N.R的三类质量感知方法，并将其应用在图像压缩&传输的不同环节，实现图像质量的感知驱动。.(2)针对“基于感知质量驱动的水声图像压缩方法研究”，提出了在感知质量驱动下的自适应测量压缩和自适应编码压缩。图像质量感知驱动的自适应策略结合图像统计先验和内容信息，动态调整压缩策略，提高图像压缩质量。.(3) 针对“基于感知质量驱动的水声图像传输方法研究”，提出了结合内容识别的图像自适应传输、掩蔽和去噪方法，在通信链路资源有限的情况下通过图像修复技术提高传输质量。