Use of BP Neural Network in Near-Infrared Spectroscopy Calibrations for Predicting of Wood Density

Pai Li; Hong Fu Zhang; Yao Xiang Li; Ya Zhao Zhang; Hui Juan Zhang

doi:10.4028/www.scientific.net/AMR.129-131.306

Paper Titles

The Synthesis and Characterization of Polyether-Modified Polysiloxane Phosphate
p.285

Effect of Annealing Atmosphere on the Mg₂Si Film Growth Deposited by Magnetron Sputtering
p.290

Preparation and Characterization of Negatively Charged Polysulfone Hollow Fiber Membrane by UV-Photografting Polymerization of AMPS
p.295

Fault Diagnosis of Rotating Machinery Based on Local Wave Decomposition and Independent Component
p.301

Use of BP Neural Network in Near-Infrared Spectroscopy Calibrations for Predicting of Wood Density
p.306

Corrosion Properties of 20^# Pipelines Steel in Near-Neutral Soils
p.312

Directly Catalytic Decomposition of H₂S to Sulfur and Hydrogen under Microwave Conditions
p.317

Numerical and Experimental Investigation into Hot Stamping of High Strength Steel Sheet for Auto B Pillar Reinforced Panel
p.322

Real-Time Registration Architecture for Far Infrared and Visible Image Using Mutual Information
p.328

HomeAdvanced Materials ResearchAdvanced Materials Research Vols. 129-131Use of BP Neural Network in Near-Infrared...

Use of BP Neural Network in Near-Infrared Spectroscopy Calibrations for Predicting of Wood Density

Abstract:

Application of BP neural network and NIRS for larch wood density prediction was investigated in this paper. The original spectra were collected and pretreated with the first derivative and 9 point smoothing. Eleven typical wave lengths were selected as BP network inputs to establish prediction model for wood density. Model was validated using cross-validation approach. The prediction correlation coefficient (R) is 0.916 while the root mean square error of prediction (RMSEP) is 0.0221. The results showed that using BP neural network in near-infrared spectroscopy calibration could significantly improve the model performance in order to rapidly and accurately predict wood density.