Monitoring machining tool wear on platform Labview using machine vision

Dérick Augusto dos Anjos de Assis¹; Joseph Kalil Khoury Junior¹; Geice Paula Villibor¹; Francisco de Assis Carvalho Pinto¹; Bruno Botelho de Souza¹; Nery Wilson Corrêa Filho¹

¹ UFV

doi:10.20906/CPS/COB-2015-1928

Resumo

Machine vision systems have been used for monitoring of the cutting tools as a preventive measure and avoid loss the premature insert. The National Instruments (NI), hardware PXI and LabView program also have been used for it. This study aimed to develop software for image acquisition and processing to identify damage and flank wear on machining tools. The program will classify two insert classes, which separates wear flank and damage, so if insert is classified like damage, it will be discarded immediately, if not the program run to measure flank wear. The development of the classification phase was used 200 flank wear and 23 damage images insert from the University of Florida (UFL). Texture analysis techniques and Bayesian method were used to classify the inserts in the two classes, damage or flank wear. The next step was measure VBmax. Sobel filter was the best solution to edge detection. The function block IMAQ Clamp Horizontal Max VI was used to measure the flank wear maximum. The image acquisition was used a Vimicro camera, PXI USB, the Vision and Motion Labview library . The system and image processing were tested with four inserts (16 images) from the Federal University of Uberlândia (UFU). Microscope image were used for validate accuracy of the system. The Kappa coefficient to evaluate the classification between the damage and the flank wear, from the UFL images, was 82.2%. The accuracy of measures Flank Wear Maximum, from UFU images, was 95.5%. The program developed in LabView platform and PXI-NI proved acceptable for acquire, process and monitor images worn insert.

Palavras-chave: Automation; Machining; flank wear