该文研究了经验模式分解(Empirical Mode Decomposition，EMD)方法在红外气体检测中的应用。针对红外气体浓度信号的特点对算法进行了改进，使得算法能够在不分解出所有IMF(Intrinsic Mode Function)分量的情况下，正确提取所需的IMF分量。与低通滤波和锁相等方法进行了比较分析，结果表明，改进后的EMD算法具有所需数据量少、计算量小、提取的有用信号能量损失少等优点，可以快速识别气体浓度的变化，有效提高气体浓度测量的精度，适用于在线实时检测。
Based on the analysis of the Empirical Mode Decomposition (EMD) method, an improved algorithm for signal processing in IR optical gas detection is presented. It decomposes only frequency-modulated signal without need to decompose every Intrinsic Mode Function (IMF) component. Compared with the methods of low-pass filter and digital lock-in amplifier, the improved algorithm has advantages of short data set, fast computation speed and less signal energy loss. It is more effective and time-saving than the conventional EMD method. The changes of gas concentration can be diagnosed by the algorithm and the precision can be improved at the same time. Thereby the gas concentration can be determined accurately and efficiently and it is suitable for real-time detection.