Model for identification of correct positioning of parts in a pick and place system

Abstract

This article investigates the use of automatic learning classification techniques applied to the task of recognizing the correct shape and color of pieces in a connector using neural networks. The system presented here shows that you can use a set of features extracted from the side view of the piece to recognize the shape of the piece and the color. The proposed model is based on two stages, one performs detection and the other is for recognition. In the first stage, color segmentation algorithms have been tested. In the second stage, a method of extracting personalized features in a color recognition approach is used. Finally, the use of a multilayer artificial neural network (ANN) is proposed to recognize and interpret the different possible shapes and colors with which the pieces can come.