摘要
针对短波通信中Morse人工接收存在劳动强度大、实时性差的问题,提出一种Morse自动检测算法。对时频图像进行图像增强和尺度归一化预处理。提取图像的纹理特征和形状特征,并将特征组合后进行高斯归一化。最后选择不同核函数设计支持向量机分类器,基于交叉验证选取最优分类器用于Morse的检测。通过实验测试,验证了算法的有效性。
To solve the problem aboutmanual receptionof Morse signals in HF communication,a new algorithm of Morse automatic detectionis proposed.In this paper,first of all,image enhancement and scale normalization are used to time-frequency images.Then,the texture feature and shape featureare extracted,and Gaussian normalizationis used tonormalize the feature combination.Finally,support vector machine classifier is designed by different kernel functions, and cross-validation is used to select the optimal classifier for Morsedetection.The experimental tests validate the efficiency of the proposed algorithm.
作者
龚智贞
袁野
孙中华
贾克斌
张海瑛
吴玲玲
GONG Zhizhen YUAN Ye SUN Zhonghua JIA Kebin ZHANG Haiying WU Lingling(College of Information and Communication Engineering, Beijing University of Technology, Beijing 100124, China The 54th Research Institute of CETC ,Shijiazhuang Hebei 050081 ,China)
出处
《无线电通信技术》
2016年第6期21-24,55,共5页
Radio Communications Technology
基金
国家自然科学基金项目(81370038)
北京市自然科学基金项目(7142012)
北京市科技新星计划(Z141101001814107)
中国博士后科学基金(2014M560032)
北京市教委面上项目(km201410005003)
北京工业大学日新人才培养计划(2013-RXL04)
北京工业大学基础研究基金(002000514312015)
