A Metal Classification System Based on Eddy Current Testing and Deep Learning

This paper proposed a novel system for classifying different metal samples based on a double-coil sensor. Traditional classification methods require a known lift-off for the sensor, which can be challenging due to the mechanical vibration of the conveyor belt. Therefore, to overcome the adverse effects of the disorderly change of lift-off, this paper innovatively combines eddy current testing (ECT) with deep learning, utilizing the nonlinear fitting ability of neural networks to distinguish five types of metals: aluminum, zinc, tin, brass, and titanium. Firstly, the paper designed a coaxial double-coil probe, which can minimize the influence of asymmetry of samples’ shape and posture. Then we constructed a driver model for the deep learning including deriving the theory of ECT and selecting categorization features by K-Means clustering. Additionally, the two-tier classification networks were created, and an impedance collector was designed to get data from the double-coil sensor, after inputting the data into our neural networks, the networks could output the classification results finally. In the tests, selecting flat-bottom metals with inclination and elevation to examine the algorithm, and using a conveyor belt with about 1 mm vibration amplitude and 0.3 m/s transmission speed. Upon comparison, it is concluded that deep learning method and cluster center features have significant gains for classification, the accuracy of this classification system can reach 94.3%.