Image Processing and Artificial Neural Network for Cam Identification

10.6567/IFToMM.14TH.WC.OS13.125

C. Pop；E.-C. Lovasz；S. M. Grigorescu；F. Pop；A. Davidescu；I. Cărăbaş

cam mechanism ； cam profile ； motion curves ； image processing ； artificial neural network

Proceedings of the 14th IFToMM World Congress

14th-4（2015 / 11 / 04）

381 - 386

英文

An automated neural network application for feature-based object recognition is presented. The main purpose is to sort mechanism pieces like cams using a new and fast method based on image processing tools. In this paper is described a cam identification application implemented in MATLAB working environment designed to determine certain specific geometric features necessary for the process of identification and classification. The idea underlying the mentioned application is to obtain the profile of the cam in order to train a two layers neural network. The network, the input and output vectors are presented.