Processamento computacional de imagens para análise de descontinuidades de soldagem em bicicletas

Abstract

Computational image processing is one of the main instruments for further investigation in microscopy, and is very recurrent in metallurgy, allowing for qualitative and quantitative study regarding the microstructure addressed. This application to welding technology has few reports of investigations, mainly for analysis on low carbon steels, which are very widespread in the metal-mechanical industry. Welding on steels of this type is present in large-scale production, such as, for example, on bicycle frames, which are presented as inexpensive and ecologically sustainable transport and in constant growth. The main type of welding for the industrial production of bicycles made by steel frames, is MIG / MAG, which through its semi-automatic welding method achieves high productivity rates, however discontinuities in the welding microstructure occur, and can harm production. The analysis by computational processing, using software such as ImageJ, allows the development of a standardized and automated method, applicable in welding microscopies, performed in steel tubes with low carbon content. Such standardization, through these computational tools of investigation of micrographs, allows the creation of plugins within an analysis program, specifically ImageJ, that is, it grants a new patentable software registration that verifies the existence and identification of the aforementioned welding discontinuities.