Implementación de un sistema de control y monitoreo en base al procedimiento de imágenes digitales en los sistemas de visión artificial aplicado al reconocimiento de la máquina selectora de botellas en el laboratorio de la Escuela de Ingeniería Industrial de la ESPOCH.

Abstract

An artificial visioning system which controls the quality based on image processing in the bottle classification machine in the lab of Industrial Engineering School at ESPOCH was implemented. It is aimed to classify three types of bottles such as: aluminum, glass and plastic through their size and shape. Light, environment, camera, digitizer card and software are the necessary parameters. MATLAB software is the best option to program the classification system by using C++ language with supported algorithms and codes with a variety of accessories as well as by its great features and facilities on driver download. Through MATLAB and its libraries, the camera selection by using weighted characteristics analysis. The Logitech of 5 Mpx webcam is the best option getting 100% of compatibility with the software which capture clear images with enough information to process 30 photos per second. Arduino Mega 2560 is used as a digitizer card, it has been chosen among several options through an analysis of characteristics and compatibility. MATLAB allows you to download and install Arduino drivers to use each PIN with algorithms in order to synchronize an intelligent automated system. The work environment and the light depend on the test object as well as the quality characteristics to be inspected. There was the necessity to use a closed rectangular structure with White background and a diffused soft light to classify three types of bottles which will focus 100 per cent on each object and in this way avoiding shadows extension. The classification process occurs through two solenoid valves that can trigger the cylinders putting each bottle at the specific locations.