【作者】 何立刚；

【导师】 张汝波；

【作者基本信息】 哈尔滨工程大学 ， 计算机应用技术， 2005， 硕士

【摘要】 本论文研究的是噪声背景下的莫尔斯信号的检测和识别技术。通讯系统中不可避免的要夹杂干扰,对于通过短波信道传输的莫尔斯信号更是如此。这对于从事人工接听莫尔斯信号的工作人员非常不利,它增加了接收人员的听觉和脑力疲劳。利用计算机实现短波信道传输的莫尔斯信号的自动检测和识别就可以减轻相关人员的劳动负担,改善工作环境。 针对莫尔斯信号的周期频率间断性的特点,本文给出了将莫尔斯信号经过短时傅立叶变换后形成的时频谱作为一幅图像、从二维的角度研究一维莫尔斯信号特征的方法。在这种思想的指导下,文中采用了数字图像处理技术,通过理论和实验分析了不同算法的优缺点和对莫尔斯信号检测的适用性。 在对莫尔斯信号谱图的图像增强过程中,主要采用了基于小波的反锐化掩模增强和基于对比度增强的算法;图像的分割主要是采用阈值分割的方法,讨论了全局阈值法和局部阈值法的增强效果,包括基于灰度均值法、基于类间最大方差的自适应阈值法和全局和局部阈值相结合的分割算法等;另外,针对莫尔斯信号谱图的特点引入了数学形态学的处理方法对噪声进一步处理,同时,给出了主要是针对目标形状的特征提取算法:最后,运用统计模式识别和智能纠错相结合的方法实现了莫尔斯信号到文本的转换。 通过理论分析和对实际含噪短波莫尔斯信号的实验,验证了本文中采用的算法的有效性,具有一定的实用价值。更多还原

【Abstract】 In this thesis the Morse telegraph code detection and recognition in noise background is investigated. It is inevitable that the communication system is disturbed by noises, especially in the short wave communication. This is great disadvantage for people who engaged in receiving signals artificially. Realizing short Morse automatic detection and recognition by computer can lighten relevant people’s burden and improve their work environment.On the basis of the essential character of Morse signal’s periodicity and frequency discontinuity, a method is put forward that Morse signal’s spectrogram through STFT(Short Time Fourier Transform ) is considered as a image and researched from the view of two-dimension. With this idea’s guidance, digital image processing technology is adapted and various algorithm’s advantage and disadvantage and their applicability for Morse signal detection are analyzed.In image enhancement, antiunsharped marking enhancement and contrast enhancement are adopted; In image segmentation, several algorithm, such as, threshold segmentation based on Arithmetic mean of Gray Value, adaptive threshold segmentation based on maximum between-cluster variance and segmentation based on combination of global and local threshold are discussed. At the same time Mathematical Morphology is introduced to deal with noises and feature abstraction algorithm mainly according to object’s shape in Morse signal spectrogram is proposed. Finally, statistical pattern recognition and intelligent correction are applied to achieve Morse signal to texts translation.Theory analysis and experimentation for Morse signal with noises shows that the whole algorithm is effective and feasible.更多还原