Research Of Stress Wave Defect Detection Of Wood Based On Wavelet Transform

The non-destructive testing of wood can accurately detect the internal defects of wood without damage to the performance of wood.It is a practical means for wood detection,which can significantly improve the utilization rate of wood.Considering the cost of testing,the safety and the operability of the experiment,this paper will study the non-destructive testing of stress wave.Although the stress wave non-destructive testing technology has already made many achievements in the study of the internal defects of wood,most of the research is still from the perspective of time domain,and judge the propagation time of the stress wave and the change of the propagation speed to diagnose whether the wood contains decay,void and other defects.This paper studies the non-destructive testing of stress wave wood from the frequency domain.Based on the study of wavelet transform,neural network and cell reverse projection imaging,the detection of wood defects and the reconstruction of wood fault images are explored.In this paper,the wavelet transform is used to process the stress wave signal,and the wavelet threshold de-noising,wavelet packet energy feature extraction and neural network are studied.The de-noising effects of different decomposition levels and different wavelet bases are discussed.The health Mongolian Oak specimens and the Mongolian Oak specimens with empty are researched by the method of wavelet packet and RBF neural network loosely bound.Wavelet transform is used as the pre-processor of the neural network,and stress wave detection signal is analyzed by 5-layer wavelet packet,using wavelet packet transform.Construct a 8-dimentional eigenvector and extract characteristic of defect energy.Then design the structure of network,train the radial basis function neural network and establish a diagnostic model,using the defect energy feature as the input of the neural network.The reverse projection algorithm is mainly researched in this paper.The inverse projection matrix is constructed by the energy and symbolic energy matrix which is extracted by wavelet packet.Then the inverse projection matrix is used to realize the reconstruction of the image of the wood fault.In addition,the quality of the reconstructed image can be improved by increasing cell density and interpolation inversion.Finally,the median filter and mean filter are used to optimize the fault image,and the fitting degree of the reconstructed image is analyzed.The combination of wavelet packet and radial basis function neural network can effectively detect the defects of wood,and the reverse projection algorithm can realize the reconstruction of the image in the frequency domain.Therefore,this research has importantly practical significance in the field of non-destructive testing of wood.