An application of sensory testing system discrimination of steel types by sparks: applying neural network

Many product inspection processes rely on the human senses, visual mostly. Generally, such inspections are referred to as sensory inspections. Recently, with the progress of data processing techniques by neuro, fuzzy, and so on, these techniques have come to be used for automating or mechanizing the sensory inspections. The paper discloses an experimental model for discriminating steel types resorting to image processing techniques and a neural network based on Method of Spark Test for Steels (JIS G0556) in the simplified test for identifying the material from different material tests executed in iron and steel fields