A novel classification method for non-magnetic metals with complex surfaces

This is the authors' abstract. We don't add key points for this paper.

Metals have extremely high recycling value, and recycling and classification of metals using eddy current sensing technology can promote the sustainable utilisation of resources. Whereas most of the current research focuses on planar metals, the study of metals with complex surfaces is still insufficient. For the latter, the distance and angle between the sensor and the measured surface are constantly changing, which will seriously affect the classification accuracy. To address this problem, this paper proposes a novel method for classification that is immune to the complex surfaces of the samples. First, the equivalent height method is proposed, by which the metal samples with complex surfaces are converted into planar samples with equivalent heights. Then, image recognition technology is used to access the height information of the sample, and eddy current sensing technology is combined to eliminate the influence of height on the classification accuracy. After experimental verification, the classification accuracy of this method is 96.67%.