Research On Grading Wood Of Chinese Zither Panel Based On Near Infrared Spectroscopy

Currently,the instrument production industry relies mainly on the subjective judgement of instrutmental technicians when selecting the wood for Chinese zither panels.However,this method lacks a summary of scientific theories and is inefficient,which limits the objectivity of the selection and the improvement of the yield.Moreover,the current model for judging the wood grade cannot satisfy the large demand of the musical instrument market.Therefore achieving rapid and intelligent grading of wood for Chinese zither panels is an urgent problem to be solved.Near-infrared spectroscopy contains information about the molecular structure of an object and is very suitable for measuring organic substances containing hydrogen.The chemical bonds of the main chemical components of wood used in Chinese zither panels are composed of hydrogen-containing groups,and the chemical compositions of the panels of different grades are different.These differences are reflected in near-infrared spectral data by light,which makes it possible to judge the wood grade.Simultaneously,convolutional neural network(CNN)has a strong feature extraction ability for non-linear data.Therefore,this paper proposes a method to analyze the spectral data by using the CNN model to determine the wood grade.The main research contents are as follows:(1)To denoise the near-infrared spectral data and reduce the amount of experimental calculations,this paper proposes to use spectral preprocessing and data compression methods.Through comparative analysis of preprocessing methods such as multivariate scattering correction,standard normal variable transformation,and second derivative derivative(Savitzky-Golay),the final preprocessing method is determined by using the smoothing effect visualization,root mean square error,and square sum of data signals as evaluation indicators.The method is Savitzky-Golay second derivative,and the best filtering window size is 15.Then,by comparing two data compression methods,improved principal component analysis and continuous projection algorithm,it was finally determined that the number of principal component variables used in the experimental analysis was 20,and the bands 1163-1244nm,1345-1375nm and 1525-1586nm may be the main characteristic bands used to distinguish different grades of wood for Chinese zither panels.(2)A wood grade discrimination method for Chinese zither panels based on Convolutional Neural Network Model is proposed.The best data dimensionality reduction method was determined through experimental analysis,and the structure and parameters of the CNN model were determined.To further optimize the performance of the CNN model,the parameter values obtained through the pre-training of the autoencoder network are used as the initial values of the parameters of the CNN model,thereby increasing the rationality of the parameter initialization.And then an optimal model CZCNN was established based on optimization strategies such as parameter regularization,exponential decay learning rate and multi-channel convolution.(3)To speed up the training rate of the CNN model and improve its classification ability,the CZCNN model is further optimized.Therefore,a wood grade discrimination method for Chinese zither panels based on the extreme learning machine optimized convolutional neural network model was proposed.The number of hidden layer neurons was determined through experiments,and then the best discriminant model CZCNN-ELM was established.To characterize the feature extraction capability of the CZCNN-ELM method,the t-distributed stochastic neighbor embedding method was used to visually compare the digital features before and after the experimental model.Finally,compared with the classic classification method,the method proposed in the paper has faster discrimination and higher recognition accuracy.In summary,the experimental results show that the proposed method can efficiently process spectral data and effectively identify the key features that distinguish different grades of wood for Chinese zither panels.While realizing the fast and accurate judgment of the wood grade of the zither panel,it also frees manpower,thus providing certain technical support for the selection of materials for the broad musical instrument market.