强噪声背景下微弱信号的提取一直是超声信号处理领域研究的一个难题,传统的信号处理方法难以准确提取弱缺陷信号,稀疏分解方法为提高超声弱缺陷的检出率开辟一条新途径,但计算量大是困扰其应用的一个主要因素.本文提出一种人工鱼群优化匹配追踪的快速算法.人工鱼是一种新型智能优化算法,具有并行寻优、全局收敛性好,对初值不敏感的特点.利用本文算法在重建信号质量不变的情况下,提高稀疏分解在冗余字典中原子匹配的速度和精度,满足信号处理实时性要求.采用与超声信号最优匹配的Gabor函数,经伸缩和平移生成过完备原子库,提高对超声信号的表达能力.通过仿真分析和实际检测铸钢试件,表明该方法能够有效地检测出强噪声背景下的弱信号.%In the field of ultrasonic signal processing, extracting weak signals under the background of strong noises remains always a difficult problem.Traditional signal processing methods are hard to successfully extract weak signals.Although sparse decomposition is a new approach in weak signal detection, its huge calculation is one of the main reasons hindering its application.In this paper a new fast algorithm is proposed, which utilizes artificial fish swarm optimized match pursuit.The artificial fish swarm algorithm is a new kind of intelligence optimization algorithm.It is advantageous of distributed parallel searching ability, good global astringency and positive insensitivity to init-values.In the circumstances of unchanging quality of signal reconstruction, the proposed algorithm improves the accuracy and increases the speed of atom matching in the redundant dictionary, and meets the requirements of real-time signal processing.The Gabor function of best matching with ultrasonic signals is adopted by the model and the redundancy dictionary is built up by stretching, compression and time shift of Gabor, thus improving the expressing ability of ultrasonic signals.Both simulated and actual cast-steel defection signals were tested.Experimental results show that weak signals can be detected efficiently out of strong noises.
展开▼
