Research on the intelligent identification of defects in ultrasonic scanning of laminated composite plates

Composite materials have excellent properties such as low cost, high strength and light weight. In recent years, they have been widely used in the aviation and aerospace industry. The advantages of ultrasonic detection, such as no pollution and easy automation, make it the most commonly used detection method in industry. However, its detection accuracy is greatly affected by the experience of detection personnel, and it is urgent to develop intelligent detection methods. Based on the above problems, six classification models including decision tree, K-nearest neighbor classification method, logistic regression, support vector machine classification, support vector machine classification based on principal component analysis, and artificial neural network were used to study the defect recognition rate of signal Fourier transform, discrete cosine transform and discrete wavelet transform respectively. The results show that in the case of the same classifier, the discrete wavelet transform can show higher recognition rate, and the artificial neural network method based on discrete wavelet transform is the best defect recognition method.