算法
降噪
粒子群优化
希尔伯特-黄变换
数学
信号(编程语言)
奇异值分解
相关系数
滤波器(信号处理)
模式识别(心理学)
白噪声
计算机科学
人工智能
统计
计算机视觉
程序设计语言
作者
Jianmin Yuan,Liangquan Jia,Lu Gao,Hengnian Qi,Xu Huang
标识
DOI:10.1109/icitbe54178.2021.00065
摘要
A SVD, EMD, SG combined algorithm was proposed to denoise the signal of TDLAS rice seed respiration detection, for seed respiration was a nonstationary procedure and sensitive to environmental changes. Firstly, the amplitude signal of the second harmonic was denoised by singular value decomposition (SVD) adaptively. Subsequently, the denoising signal was decomposed by Empirical mode decomposition (EMD). The dispersion entropy (DE) of each IMF and the correlation coefficients between each IMF and the denoising signal were calculated as the common coefficient, which was used to determine the effective component. Finally, Savitzky- Golay (SG) filter was used to smooth the signal, the polynomial order and window size of SG filters were obtained by improving particle swarm optimization (PSO) algorithm. A simulated signal and experiment data acquired from the TDLAS system were processed, the denoising results were compared with those from the traditional algorithms, this method showed better performance with the higher signal-to- noise ratio and correlation coefficient. The SVD, EMD, SG combination algorithm was suitable for TDLAS rice seed respiration detection
科研通智能强力驱动
Strongly Powered by AbleSci AI