Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
ASSESSMENT OF CUTTING TOOL WEAR USING COMPUTER VISION
Submission Author:
Breno Costa , MG
Co-Authors:
Breno Costa, Antônio Maia, Alexandre Abrao, pedro bastos costa
Presenter: Breno Costa
doi://10.26678/ABCM.COBEM2021.COB2021-1465
Abstract
In the traditional approach to determine tool wear in machining operations, the cutting tool is removed from the machine tool and tool wear is evaluated by the operator. This procedure demands time and human intervention, and the results can be significantly influenced by the operator. An erroneous evaluation may lead to wrong decisions that can compromise the quality and efficiency of the manufacturing process. An alternative is to evaluate tool wear inderrectly during the manufacturing process by using image metrology. This procedure can be automated using algorithms and systems for image acquisition and analysis, thus providing a fast method for wear measurement using a standardized procedure. This work aims to present the development of a computer vision algorithm to evaluate the wear of cutting tools through image analyses. The procedure for tool wear measurement encompasses image acquisition and recognition, morphological transformations, edge detecting, and wear measurement. In this work, the acquired image with both a reference object and the worn cutting tool is processed by the software, firstly determining the number of pixels that the width of a reference object occupies in the image, since this width is known, a scale factor relating pixels to linear length can be determined. The next step is obtaining the number of pixels containing the wear land of the cutting tool, using the scale factor the pixels are converted to linear length. Preliminary results indicate that the maximum flank wear measured with this method presented an accuracy from 0,99% to 4,90%. It was also noted the feasibility of using a unique scale factor, avoiding the need of reference objects for each image extraction. It was noticed that a standard scale factor is effective for different measuring experiments since the tool is located at the same position and the images are acquired preserving the same conditions such as distance of the tool from the camera and lighting. The data provided by the system can be used to expand future research in the field of machining, avoid machine and tool holder damage caused by catastrophic tool failure, determine the time to replace cutting tools with better accuracy, reduce rejected parts and increase productivity.
Keywords
Computer Vision, image metrology, measurement device, cutting tool wear, wear measurement
