一种优化的舰船噪声连续谱特征提取算法研究  

An Optimized Algorithm for Continuous Spectrum Extraction of Ship Noise

作　　者：姚运启[1] 张效民[1] 赵延安[1] 

机构地区：[1]西北工业大学航海学院,陕西西安710072

出　　处：《计算机仿真》2011年第5期5-8,共4页Computer Simulation

摘　　要：研究舰船辐射噪声,是目标识别的依据。连续谱是水下目标检测和识别的重要判别特征,为在目标出现后最短时间完成判决,要对相应目标特征参数进行提取,提出了更高的时效性,并能实时快速提取。针对传统算法运算周期长和段连接处有毛刺的干扰问题,采用一种快速且有效算法。用分段和固定拟合频率的结合,再在段与段连接处重叠点的基础上采用直线代替曲线,可有效降低程序运行时间,解决干扰问题。对实测数据进行仿真,结果表明,新算法比传统方法更有时效性,提高了水中检测的实时性。The continuous spectrum future is one of the important distinguishing criterions for underwater target detection and identification.The corresponding target feature extraction requires timeliness for completing the sentence in short time when the target appears.Because of the problem that it has a long computing cycle and has the bur on the cohesion department for the traditional algorithm,the paper presents a fast and effective algorithm.First,it uses segmentation combination of the fixed-fitting frequency,and then uses a straight line instead of curves on the cohesion department based on the segment overlap junction points,which can effectively reduce the run time and solve the glitch problem.The simulation of measured data can get the desired result.Compared with the traditional method,the algorithm is more time-sensitive and has practical value in the practical project.

关 键 词：分段 固定拟合频率 直线代替曲线 连续谱提取 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]