Abstract: In view of the wood component glue line defect, a method of wood structure nondestructive detection was proposed based on ant colony BP neural network. The wood specimens was tested to obtain the test signal by ultrasonic testing instrument, in order to eliminate the testing effect of the tester gain control and defect size, angle variation on the test defect echo amplitude, the defect signal amplitude was needed to normalization. The wood component decomposition of ultrasonic signals was de-composite to different frequency channels by the domain band-pass characteristics of the wavelet frequency. By extract characteristic of the original signal in different frequency channels, the ant colony neural network could train the parameters and examine the position of the wood components with defection. The test results show the effectiveness of the proposed method.

Key words: wood structure, wood component glue line defect, wavelet transform, Ant-colony BP network