The Research Of The Monitoring Technology For The Multi-component Gas Based On FTIR

Enviromental issue has been the concern at home and aboard .With the development of the modern industry, the environmental pollution problems is becoming more and more serious, so the establishment of online realtime remote environmental monitoring system is imminent. While in many developed countries aboard have been designed many of the toxic and harmful gas remote sensing monitoring system,and has been used in environmental monitoring ,but the technology is still at experimental stage in our country .Enviromental monitoring instruments made in china still have many disadvantages .With unique advantages, The FTIR is widely used in air monitoring in indoor and outdoor.This paper designed an online real-time environmental monitoring system based on the FTIR which is dual processor ARM and DSP.The wavelet neural network is applied to processing the spectrum signal of muti-component gas. the main works and achievements are as follows :(1) Intruduced the principle of the FTIR spectroscopy , analysed the qualititive and quantitative monitoring principle of FTIR spectroscopy.(2) Designed the singal processor of multi-component gas monitoring system based on the technology of FTIR . Completed the devise of ARM, included the hardware and software.In the hardware , we designed the key circult module , and the HPI is the communication interface of ARM and DSP .whle in the software , Completed the establishment of Linux cross-compiler environment,the transplantation of the ARM ( s3c2410X ) platform Linux kernel and boot_loader cutting on the host (PC) platform.(3)This paper presents a new method based on wavelet neural networks which can achieve the simultaneous removal the noise and baseline drift existing in the spectrum singal .constructed different wavelet function as the neurous in neural networks. Comparaed through experiments ,the wavelel neural networks can remove the noise and baseline drift simultaneously with better effect than other methods.Give preliminary study on qualitive and quantitative in the multi-component gas with wavelet neural networks .