NON-CONTACT METHOD FOR METAL RECOGNITION

The present article proposes a non-contact method for metal recognition - aluminium, chrome-nickel steel, brass and copper. A piezoelectric emitter sends an acoustic wave to the object to be identified. The reflected signal is received and processed using orthogonal wavelet basis functions. Clusters are formed based on feature spaces of classifiers realized following the knearest neighbour (kNN) method. During each subsequent measurement the results are related to a specific cluster. This approach has yielded a 100% recognition of the above materials, as well as three types of explosive