Apparatus for detecting vibrations of a test object using a competitive learning neural network in determining frequency characteristics generated

摘要

There is known a testing method of inspecting whether a building wall is delaminated or peeled off by aurally listening to the sound that generated by hammering the wall. Moreover, there is a widely known method of listening to the sound made by tapping on produce to check whether a fruit such as watermelon is ripe. Such methods of estimating the internal state of a test object by striking the test object to impart vibrations thereto and hearing sound generated by the vibrations with one's ears is widely applied in various fields. There is also a widely known technology of ultrasonic diagnosis, which examines an internal state of a test object by emitting acoustic waves, for example, ultrasonic waves, onto the test object to impart acoustic vibrations therein and analyze the acoustic vibrations reflected from or transmitted through the inside of the test object. However, in the method of estimating the internal state of the test object by hearing the sound generated when striking the object, only an experienced person can assess the state by the sound, so only a limited number of persons can take part in the inspection.