Study On Detection Of Moisture Content Of Wood Based On Near Infrared Spectroscopy

Moisture content of wood is an important characteristic parameter, real-time、 nondestructive and accurate detect moisture content of wood has an important role in promoting the wood processing industry productivity and technology upgrading.In view of the above problems that drying method and electric measurement had, this paper proposes a method of wood moisture content detection based on near infrared spectroscopy. The technology is a kind of detection technology with short response time, high measurement accuracy, and can achieve non-destructive testing and other advantages. The research of this topic has practical significance for the detection of moisture content of wood.This paper describes near-infrared absorption theory and near infrared spectral properties of water, introduced Beer-Lambert Law and the choice of direct absorption method to detect moisture content, Based on describes the basic principles of the direct absorption of technology and the international general HITRAN spectral database, the simulation draws the H2O absorption spectrum and other interference gases of the spectral line. Take into account the interference of laser performance and the strength of the background gas, its absorption spectrum analysis to determine 1899.703nm wavelength as the characteristic absorption peak. The method of reference wavelength is used to eliminate the influence of the surface state of the measured object. Finally, the wavelength of 1896.136nm and 1902.686nm is selected as the reference band. Designed and built based on near-infrared spectroscopy wood moisture detection system; using DFB lasers and laser controller for laser wavelength modulation; aiming at the problem that the laser beam is divergent and irregular, the laser filter is proposed to improve the laser beam profile; complete signal acquisition, preprocessing, storage and waveform display and other functions, based on STM32 and Lab VIEW platform, realize the automation of the testing process.This paper built a system of timber moisture detection based on the near infrared spectroscopy rate, the first test is the performance of the system. By taking tunable laser optical power and output wavelength stability tests to demonstrate the stability of the near-infrared light; the DFB laser tuning characteristics tested to determine its operating characteristics. Range of moisture near the fiber saturation point is detected, the moisture content range for the blind conductance method, according to experimental data, a relatively linear least squares method of fitting, quadratic and linear fitting of artificial neural network data fitting method, choose artificial neural network to obtain a linear relationship between the signal amplitude ratio and moisture content of wood, the correlation coefficient is 0.9841, root mean square error is 0.0167; near infrared moisture content detection system detect by 3 groups of test materials moisture content, and the results prove that the moisture content of the measured value and the value of the aqueous outspoken same trend, both the absolute error is less than 0.6%; finally, through the outer layer of the internal moisture content of wood and different wood grain moisture content were detected compared initially proved the feasibility of measuring moisture content of wood technology based on near-infrared spectroscopy.