CAJ | 학술논문

超声导波在高速公路护栏立柱检测中存在信噪比低、回波中特征信号不明显等问题,为此,提出一种改进的子空间匹配追踪算法( ISMP),利用回波信号的先验信息在过完备Chirp原子库上得到每次迭代的强相关原子集,经过迭代得到待匹配信号的最佳时频原子,从而实现对立柱回波信号的特征提取。通过对中心频率为128 kHz的检测信号进行算法验证,结果表明,ISMP可以有效提取出回波信号的特征原子,所得检测长度与实际测量误差小于1%,满足工程检测要求。
초성도파재고속공로호란립주검측중존재신조비저、회파중특정신호불명현등문제,위차,제출일충개진적자공간필배추종산법( ISMP),이용회파신호적선험신식재과완비Chirp원자고상득도매차질대적강상관원자집,경과질대득도대필배신호적최가시빈원자,종이실현대립주회파신호적특정제취。통과대중심빈솔위128 kHz적검측신호진행산법험증,결과표명,ISMP가이유효제취출회파신호적특정원자,소득검측장도여실제측량오차소우1%,만족공정검측요구。
For the low signal to noise ratio, unconspicuous echo characteristic signal and other issues encountered in ultrasonic guided waves inspection of highway guardrail posts,an Improved Subspace Matching Pursuit( ISMP) algorithm is presented,which obtains the correlated atomic sets for each iteration from over-complete chirp atom library by the priori information of echo signal and achieves echo signal feature extraction by iterating to get the best time-frequency atom of the signal to be matched. The algorithm is verified by the guided wave inspection signal whose center frequency is 128 kHz. The results show that ISMP can effectively extract the characteristic atoms from echo signal, and the actual measurement error is less than 1% . It can meet the requirements of project detection.