Data management


@imveurope


www.imveurope.com


fluctuations. In his view, such an approach could be particularly beneficial because, although changes in production quality oſten have different causes, machine vision systems tend to only recognise results expressed by poorer quality. ‘It could also be beneficial to compare machine


vision data of one and the same workpiece from different production steps with each other. As a result of this comparison, previous inspection steps may be improved. Improved inspection procedures make the production more efficient, as rejects can be identified earlier,’ he said. Moreover, if such combined data sets


Trend of data generated from camera images showing a statistical histogram, statistical limits and upper and lower specification limits


detail about what was evaluated, perhaps in terms of variables like limits, performance and results. In such a way, he pointed out that with modern machine vision devices, the image is more an artefact of a process than the true data of that process. ‘Te real data provided by vision systems is the


scalars, arrays and strings that are associated with the image. On these, evaluations are applied and results generated. Today, we look to reveal and visualise the data generated in the vision systems and trend it to verify that a manufacturing process remains within specification, part aſter part,’ he said. One example that Chabot cited is the use of


machine vision to evaluate dispensing a sealant bead on a part. Taking the image as the starting point, he explained that a vision system can break down the bead into thousands of regions, and for each one measure the width of the bead and the offset from the expected centre position. ‘Each region has unique measurements


and limits used to determine whether or not the [sealant bead] in that region is acceptable. Historically, manufacturers would have been content with the overall status for that dispense action on that part,’ he said. ‘Now, they want individual measurements


because they can use these data arrays to trend the behaviour of the robot as it applies the sealant, and spot early any indicators of misalignment, worn or damaged nozzles, or any other issue that could affect bead quality,’ he added.


Improvement strategies Elsewhere, Johannes Hiltner, product manager for Halcon at MVTec Soſtware, highlighted the fact that, in general, much of the data collected by


machine vision systems is used for quality control – which commonly means rejecting defective parts or parts that do not meet quality standards. ‘Meaningfully, these quality checks are


conducted as early as possible in the production process. Te data can also be used to analyse the efficiency or productivity of production. For example, to determine how much scrap production generates, or at what stage quality problems occur through the process,’ he said. ‘Some production steps are also checked by


machine vision in addition to human inspection. Tis generates data that compares machine and human quality, and thus helps to verify the quality of a machine vision implementation in complex applications,’ he added. In addition to evaluating the data collected


from machine vision devices, Hiltner believes that manufacturers should also use it with other data, for example, read-outs related to, say, the temperature of a machine, or the electric current


determine, for example, that specific defects result from previous handling and processing activities, Hiltner pointed out that ongoing production processes could be modified as well. One example he cited is that grabbers could be soſtened and machines adjusted accordingly. ‘Of course, rising defect rates could also


be an indicator for needed maintenance of the machines employed preliminarily to the inspection process. In such a way, machine vision data can also be utilised for predictive maintenance,’ he added.


Cross-process analytics In the historical sense, Chabot highlighted the fact that data collection was always a granular process. Put simply, if users stored images from a vision system, they generally put them in a vision data image storage location, which tended to generate a number of data silos around the plant. In his view, such a limited approach also made it difficult to carry out any cross analytics – or even see efficiently what he described as the big picture of everything that happens to a part across all the various different processes, be they vision, press, leak or rundown, and so on.


Graph of the width of a temperature vulcanising (RTV) bead with part of it outside specification 30 Imaging and Machine Vision Europe • Yearbook 2018/2019


Sciemetric Instruments


Sciemetric Instruments


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60