论文部分内容阅读
光谱吸收法是对甲烷浓度检测的一种有效手段,通过棱镜气室结合光子晶体光纤的应用实现光谱吸收法对甲烷浓度的高精度在线检测。但在检测过程中,由于环境中温度、压强以及系统本身设备的影响,使得接收的信号中包含大量的噪声。支持向量机(SVM)具有泛化能力强和寻求全局最优点的特点,被用于甲烷浓度检测的信号处理。Matlab实验结果表明,使用SVM原理滤波能有效地滤除噪声,把有效的信号分离出来,并用信噪比评估去噪效果。使用该方法滤波能够使信噪比达到130dB以上,与传统的小波降噪相比有很大的提高,能达到理想的去噪目的。
Spectral absorption method is an effective measure for methane concentration detection. Through the application of prism gas chamber combined with photonic crystal fiber, high-precision on-line detection of methane concentration by spectral absorption method is realized. But in the course of testing, because of the temperature in the environment, pressure and the influence of the apparatus of the system itself, make the received signal contain a lot of noises. Support Vector Machine (SVM) is characterized by its generalization ability and global optimization. It is used for signal processing of methane concentration detection. The experimental results of Matlab show that using SVM principle filtering can effectively filter out the noise, separate the effective signal, and evaluate the denoising effect by SNR. Using this method, the signal-to-noise ratio can reach more than 130dB, which is greatly improved compared with the traditional wavelet denoising, so as to achieve the ideal denoising purpose.