基于同步提取变换和脊线检测的空间锥体目标微多普勒特征提取方法

Micro-Doppler feature extraction method for space coningtarget based on synchro-extracting transform and ridge detection

针对空间锥体目标微多普勒(micro-Doppler,m-D)提取时在m-D曲线交叉处易出现跟踪错误的问题,提出一种基于同步提取变换(synchro-extracting transform,SET)和脊线检测的m-D特征提取方法。首先,该方法利用SET技术得到能量分布更加集中的时频像,并使用Hough变换提取锥体顶部产生的正弦m-D曲线;之后基于CLEAN思想,将提取到的顶部m-D曲线从时频像中消除;最后利用基于改进疯爬算法和卡尔曼滤波的脊线检测算法提取锥体底部产生的非正弦m-D曲线。仿真结果表明,该方法可以避免m-D曲线交叉处的跟踪错误问题,正确提取m-D特征。

To solve the problem of the tracking mistakes at curve intersections of micro-Doppler(m-D)in m-D extraction of the space coning target,an m-D feature extraction method based on synchro-extracting transform(SET)and ridge detection is proposed.Firstly,the algorithm obtains the energy-concentrated time-frequency image through SET technology.Secondly,the sinusoidal m-D curve generated at the top of the coning target can be extracted by the Hough transform.Then,the extracted top m-D curve is subtracted from the time-frequency image based on the CLEAN idea.Finally,the non-sinusoidal m-D curve at the bottom of the coning target can be extracted by the ridge detection algorithm based on the improved crazy climber algorithm and Kalman filter.Simulation results show that the proposed algorithm can avoid the problem of tracking mistakes at the intersection of m-D curves,and then accurately extract the m-D features.