A machine vision system for automatic sieve calibration

Abstract

Low measurement efficiency is a recurrent problem in calibration laboratories, especially when the instruments to be calibrated are mechanical and do not have a built-in communication interface. In the case of a sieve calibration procedure, for example, low measurement efficiency caused by a small number of examined openings could compromise the quality of the judging process. In addition, conventional manual calibration methods present some intrinsic disadvantages, such as the high consumption of time and the fact that the calibration may be subject to human error. In this paper, a machine vision system (a set of hardware and software) is presented for the automatic calibration of sieves. Experiments were carried out to validate the method and equipment presented in this work and the results indicated that the proposed machine vision can measure sieves, with a high sampling of openings, and with high efficiency and accuracy. In addition, the results showed a reduction of approximately 92% in the time spent on the calibration process, when compared to traditional methods, with the same accuracy.

Surface defect detection of machined parts based on machining texture direction

“…The visual inspection technology used in the industry mainly includes machine vision methods based on artificial intelligence [7,8] and machine vision methods based on manual feature extraction [9]. To obtain better stability, this paper uses a machine vision method based on manual feature extraction.…”