基于压缩感知的天气雷达回波信号处理

To improve the detection capability and data transmission efficiency of CINRAD, to enhance the CINRAD ability to obtain information, the project proposes a novel signal processing method of CINRAD based on compressed sensing theory.. The projec researches the sparsity of CINRAD echo signal based on the radar echo chararcteristics. It accomplishes compressive sampling and recovery of signal based on the sparsity of radar echo signal, the randomicity of measurement matrix and nonlinearized optimization. As a new signal precessing theory used in CINRAD, CS provides great possibilities for overcoming inherent limitations of traditional weather radar, and has potential to resolve many problems associated with high resolutin radar, such as high sampling rate, too many data and difficulties of real time processing,we can see large potential of the theory in simplifying radar hardware,conquering data lmlitations, impoving radar detection capabilities. The research will promote the technology of the CINRAD signal processing, improve its ability of detection. The research has an important application value in detecting and forecasting severe weather.The project could be developed to provide theory basis and technique methods for super resolution of weather radar, improve detection capability greatly, enhance compression ratio of weather radar, and solve the data overlord in distributed weather radar..

为提高新一代天气雷达系统的探测能力和数据传输效率，丰富天气雷达的信息获取能力，本项目提出基于压缩感知理论的新一代天气雷达信号处理方法。. 通过对天气雷达回波特性的研究，实现天气雷达回波的稀疏化。基于雷达回波信号的稀疏性，使用随机测量矩阵和非线性算法对天气雷达回波压缩测量和重建。这种全新的信号处理理论克服了传统天气雷达的固有缺陷，可以解决天气雷达分辨率提高面临的高采样率、大数据量和实时处理困难等问题，在简化雷达硬件设计、弥补雷达数据缺陷，改善雷达探测能力等方面有巨大潜力。其成果将促进CINRAD数据处理技术的发展，提高探测能力，对灾害性天气的预警、预报有重要的应用价值。开展本项目的研究，有望为提高天气雷达的探测能力提供理论基础和技术方法，使雷达数据分辨率得到大大改善；有望提高天气雷达回波数据的压缩率，解决分布式天气雷达通信数据量过大的问题。

由于Nyquist采样定理的限制，高分辨率气象雷达面临采样率过高、数据存储量过大等问题。压缩感知理论可以实现气象雷达信号的压缩采样，解决采样率过高等问题。基于压缩感知理论，分析了气象雷达回波信号的稀疏性，建立了气象雷达回波信号的压缩采样和重建的过程。研究了的观测矩阵的优化算法，设计确定性的测量矩阵以替代难以硬件实现的随机测量矩阵。通过对测量矩阵进行优化，降低重建的采样率，提出了测量矩阵奇异值分解优化算法，提高了测量矩阵的性能。随着探测能力的提高，天气雷达数据量也、急剧增长。天气雷达数据压缩算法可以减少传输和存储的数据量，但通用的数据压缩算法并未充分考虑天气雷达数据的特点。项目提出基于径向预测的天气雷达数据无损压缩算法。分别研究了我国应用广泛的CINRAD数据和高分辨率的双极化CHILL雷达数据的特点，分别提出了适用于不同雷达的无损数据压缩算法。当前国内外气象雷达数据压缩的研究主要集中于反射率数据，本项目提出几种压缩算法适用于雷达的各种基数据产品，有较好的应用前景和理论价值。