Controle sensorial para detectar descontinuidade de cola em madeiras para fabricação de lápis

Abstract

This project consists of the design and implementation of a system capable of detecting the discontinuity of glue in wood on pencils manufacturing by utilizing optical sensors, whose objective is the application of theoretical concepts that were seen during graduation, in order to guarantee a better quality of the final product. The idea of this project emerged from an extracurricular internship at Faber-Castell in Germany. In this project the process was simulated. The images were taken by a cell phone and processed / analyzed by MATLAB. This project also includes the design of a prototype attempting to create a scenario similar to the company and to improve the whole project. The image processing in the software consists of transforming the original image into grayscale, which gives an uniform binary image. From the binary image, matrix with logic level 1 and 0, the detection of the irregularity is generated from a comparison with the sum at each pixel interval of each channel. The process feedback is generated by MATLAB itself, where each channel has its own analysis. In order to provide image capture of a moving object, a study was also developed and indicated equipments that could be used in future projects. In the course of this project, the image processing technique addressed to the problem is reviewed as well as theories and instruments for the development of this project.