摘要
分析了快速傅里叶变换(FFT)在光频域反射仪(OFDR)频谱分析中应用的局限性,论述了快速傅里叶变换-分段Chirp-Z变换(FFT-SCZT)算法的不足。提出了一种基于改进的遗传算法(IGA)和Zoom FFT(ZFFT)的快速高精度频谱分析(FFT-IGA-ZFFT)算法。描述了该算法的计算过程,并推导了算法的时间复杂度。研究结果表明,在同一运算平台下,处理10~7个OFDR采样数据时,FFT算法耗时3.130s,FFT-SCZT算法耗时1.993s,而FFT-IGA-ZFFT算法仅耗时0.525s即可获得同等精度。FFT-IGA-ZFFT算法在处理速度上具有明显的优势。
We analyze the limitation of the fast Fourier transform (FFT) method applied to the spectrum analysis of the optical frequency domain reflectometry (OFDR). The shortcomings of the fast Fourier transform-segmented Chirp-Z transform (FFT-SCZT) algorithm are discussed. Based on an improved genetic algorithm (IGA) and Zoom FFT (ZFFT), a fast and high-resolution spectrum analysis algorithm, FFT-IGA-ZFFT algorithm, is proposed. The computation procedures of the proposed algorithm are described and its time complexity is deduced. The experimental result shows that the FFT algorithm costs 3. 130 s, the FFT-SCZT algorithm costs 1. 993 s, and the FFT-IGA-ZFFT algorithm costs 0.525 s when 10^7 OFDR sampling data is processed in the same computing platform and with the same resolution. The FFT-IGA-ZFFT algorithm is outstanding in term of processing speed.
出处
《光学学报》
EI
CAS
CSCD
北大核心
2017年第4期72-78,共7页
Acta Optica Sinica
基金
国家自然科学杰出青年基金(61225004)
关键词
光通信
频谱分析
遗传算法
高精度
时间复杂度
optical communications
spectrum analysis
genetic algorithm
high resolution
time complexity