【摘要】 针对多分量线性调频(linear frequency modulation,LFM)雷达信号检测和参数估计精度低、计算速度慢等问题,提出了一种基于小波变换的切割聚类拟合参数估计的算法。该方法首先通过小波变换得到信号的三维时频分布图,其次采用等高线截取提取出小波脊线,再找出脊线的交点,以交点为界对小波脊线图进行切割,利用模糊C均值聚类完成各LFM分量脊线的聚类,最后分别对每段脊线进行拟合加权,从而估计出多分量LFM信号参数。仿真结果表明,与基于Hough变换检测直线方法相比,不仅在计算复杂度以及参数估计的准确度上都有较大的提升,而且当LFM信号分量达到4个以上亦有较准确的检测精度。更多还原

【Abstract】 To solve the problems of low precision and slow computation speed in the multi-component linear frequency modulation(LFM)radar signals detection and parameter estimation,a novel parameter estimation algorithm with segmented clustering and fitting based on wavelet transform is proposed.Firstly,the 3Dtime-frequency distribution image of the signals is obtained by the wavelet transform,and extracts the wavelet ridge of the signals with the contour interception algorithm.Then find the intersection of the ridge,and cut the image of the wavelet ridge based on the intersection.After that,all the ridges of the LFM signals are clustered by FCM algorithm.Finally,each segment is fitted and weighted respectively and the parameters of the multi-component LFM signals are estimated.Simulation results show that the computational complexity and the precision of the parameter estimation notonly have greatly improved,but also have more accurate precision under the condition that LFM components reach more than four when compared with the traditional Hough transform for the line detection.更多还原