Title :
A new method for object identification in ultrasonic systems using neural nets
Author :
Serrano, I. ; Lázaro, A. ; Oria, J.P.
Author_Institution :
Dept. of Electron. Technol. & Autom. Syst., Cantabria Univ., Santander, Spain
Abstract :
This article describes an ultrasonic identification method which uses an array of sensors to recognise an object. The special feature of this particular set-up is that the object is located on a rotating platform in front of the array. In this way, taking advantage of the signal given by the ultrasonic echo from an orthogonal face and solving the typical problems associated with acoustic dispersion, useful information can be obtained for subsequent treatment. Very good results can then be obtained using artificial neural network techniques for data processing. This method not only permits the correct identification of the objects, but also facilitates their three dimensional reconstruction
Keywords :
backpropagation; echo; feedforward neural nets; image reconstruction; object recognition; ultrasonic imaging; 3D reconstruction; acoustic dispersion; backpropagation; feedforward neural nets; object recognition; ultrasonic echo; ultrasonic systems; Acoustic diffraction; Acoustic emission; Acoustic reflection; Mirrors; Neural networks; Sensor arrays; Sensor phenomena and characterization; Shape; Solids; Surface acoustic waves;
Conference_Titel :
Computational Intelligence in Robotics and Automation, 1997. CIRA'97., Proceedings., 1997 IEEE International Symposium on
Conference_Location :
Monterey, CA
Print_ISBN :
0-8186-8138-1
DOI :
10.1109/CIRA.1997.613852