3-D object recognition using an ultrasonic sensor array and neural networks

Abstract

3-D object recognition which is independent of translation and rotation using an ultrasonic sensor array, invariant moment vectors, and neural network is presented. With invariant moment vectors of the acquired 16x8 pixel data of square, rectangular, cylindrical, and regular triangular blocks, SOFM (Self Organizing Feature Map) neural network can classify 3-D objects. Invariant moment vectors are constants independent of translation and rotation. The experimental results of the 3-D object recognition using an ultra sensor array are presented to show the effectiveness of the proposed algorithm. © 2007 Springer-Verlag Berlin Heidelberg.