基于改进进化规划的正交匹配追踪超声导波信号识别

Guided Wave Signal Recognition by Orthogonal Matching Pursuit Based on Modified Evolutionary Programming Algorithm

提出基于t算子的改进的进化规划(MtEP),进行了性能测试,发现其以更高的概率收敛到全局最优。采用chirplet函数作为匹配原子,将其应用于正交匹配追踪方法(OMP)。利用导波检测仪对含缺陷铝管进行导波检测,利用MtEP实现的正交匹配追踪方法对检测信号进行匹配分解与重构,发现重构所得导波信号信噪比明显提高。将匹配结果与未改进的进化规划算法(tEP)实现的正交匹配追踪匹配同样的信号所得结果作比较,基于MtEP的OMP比基于tEP的OMP匹配所得参数能够更加准确地反映信号激励频率,且耗时更少,因此是一种有效的导波信号识别方法。

A kind of modified evolutionary programming using mutations based on the t probability distribution（MtEP）is proposed and applied to the orthogonal matching pursuit method.Three kinds of mutation operators are compared and the chirplet function is chosen as matching atom.An aluminum pipe with holes is inspected by guided wave testing system and the measured signal is decomposed and reconstructed by the orthogonal matching pursuit method.The SNR of the processed signal is improved obviously.The matched result is compared to the matched result from OMP with EP using mutations based on the t probability distribution（tEP）.The matched parameters get from OMP with MtEP and tEP are compared and analyzed.The center frequency of the excitation signal are more exactly and the computation time is shorter by OMP based on MtEP than based on tEP.Therefore,OMP based on MtEP is a useful signal recognition approach for pipes guided wave NDT.