摘要
舰艇在进行鱼雷试验和作战使用过程中,需要以气象传真图作为鱼雷航行和作战的水文气象保障,如何使气象传真图更加清晰可读成为急需解决的现实问题。提出了一种基于信号稀疏表示的针对海上气象传真图增强的新方法。该方法以气象传真图底图和固定的气象符号作为字典函数,充分利用气象传真图的稀疏特性,采用匹配追踪算法进行气象传真图稀疏表示,并利用气象传真图的稀疏系数有效分离噪声。试验结果表明,在气象传真图增强中,该方法明显优于目前的图像增强方法。
In the process of torpedo run test or operational application of a naval vessel, we need a hydrometeorological support based on received meteorological facsimile maps, thus legible meteorological facsimile maps is necessary. A new noise removal algorithm for meteorological facsimile maps with zero-mean Gaussian noise is proposed in this paper based on signal sparse representation. This algorithm makes full use of self features of meteorological facsimile maps, takes meteorological facsimile base maps as dictionary functions, and decomposes these meteorological facsimile maps via the dictionary functions and matching pursuit to separate the information and noise. Experimental results show the superiority in performance over the existing methods for meteorological facsimile map enhancement.
出处
《鱼雷技术》
2011年第4期312-317,共6页
Torpedo Technology
基金
国家自然科学基金广义测不准原理及其应用研究(61002052)
关键词
气象传真图
稀疏表示
图像增强
字典函数
meteorological facsimile map
sparse representation
image enhancement
dictionary function
