为降低电路及光路噪声等对痕量气体检测精度的影响，达到进一步提高痕量气体检测系统精度的目的，选用超窄带激光器作为系统光源并以谐波检测原理及改进希尔伯特黄变换（IHHT）滤波算法作为数据处理的核心算法构建高精度痕量气体检测系统平台。选取甲烷作为被检测对象，理论计算得当光源输出中心波长为1650.959 nm时，该系统甲烷气体体积分数C与输出光强一、二次谐波比值I2f/If的相关系数为0.084639。将一定体积分数的甲烷气体通入平台光纤气室中进行灵敏度实验，实验结果表明应用IHHT降噪处理后C与I2f/If的相关系数由使用傅里叶变换(FFT)处理的0.062585提升到0.074884，且气体体积分数越低，IHHT的降噪效果越好。

In order to reduce the effect of electric circuit and optical path noise on the trace gas detection and achieve the purpose that further improving the accuracy of trace gas detection system, ultra-narrow-bandwidth laser is utilized as the system light source, and harmonic detection and improved Hilbert-Huang transform (IHHT) filtering algorithm are applied as the key algorithms of the data processing to construct the high precision trace gas detection system. Methane is used as the detect object, and theoretical calculation result shows that when the output center wavelength of the system light is 1650.959 nm, the correlation coefficient between methane gas volume concentration C and the first, second harmonic generation ratio of the output intensity I2f/If is 0.084639. The sensitivity experiment is carried out by inject specific fraction volume methane into the platform optical gas cell. The experimental data show that after the application of IHHT to reduce noise, the correlation coefficient between methane C and I2f/If is upgraded from 0.062585 which is obtained by applying fast Fourier transform (FFT) to process the data to 0.074884. The better effect of improved IHHT noise reduction depends on the less methane gas volume fraction.