Data management ‘Te trend now is to get away from silos and


bring data together to have better analytics – and from that better manufacturing performance by being able to target the areas where efficiency can be improved,’ he said. To help in illustrating this point, Chabot


expanded on his previous example relating to the use of machine vision technology to generate a trend for a sealant dispenser. In an effort to complement this data set he explained that he and his team also gathered data from DC tools running down fasteners – by assembling this part with another one – and finally by adding data from a leak test on the assembled part. ‘By having data from these three distinct


processes in one place, you can now start to see how results at the dispenser affect the leak test and vice versa,’ he said. ‘Merging vision data with other data generated


on the plant floor helps you maximise your ability to analyse how different processes can impact each other and help you improve yield through cross- process analytics,’ he added.


Size isn’t everything In Chabot’s view, once manufacturers eliminate such data silos, the process of managing and using any vision data collected becomes easier – largely because they can use the same tools for all their manufacturing data. In essence, this results in the establishment of a common method to manage the data and a common method to visualise the data, which greatly reduces learning curves for people interacting with these systems. ‘Of course, that means you must adopt an


Vision data can be combined with other data regarding the temperature of a machine, or the electric current fluctuations, for instance


quantities of data, Hiltner also stressed that big data and data storage are not really a big deal anymore. Tat said, he admitted that other challenges relating to standards and data consistency still exist. As he explained, standards for the


There is a trend building towards artificial intelligence taking on [analytic tasks]


transmission, storage, and interpretation of data are needed, because every device in a production line has different data structures, oſten depending on the specific manufacturer. ‘One promising approach


for this is OPC UA [Unified Architecture],’ he said. ‘Consistent and distinctive classification and interpretation of data is the even bigger


analytics platform that allows you to bring all this data together in a single place,’ he said. Chabot also highlighted the somewhat unique


challenge with images relating specifically to their size. As he explained, raw image files, which are the common output from vision systems, are quite large. However, such files are only needed if manufacturers plan on reprocessing them in the vision systems. If they only want images for compliance, he said there are many techniques to reduce image size, from changing their resolution to going to a compressed format. ‘Most of the time, you can have an image a


tenth the size of the original and the naked eye can barely notice a difference from the original image. Newer compression algorithms allow for even further space savings without compromising the quality of the images,’ he said. In terms of the challenges of managing large


challenge. Also, as far as sustainability is concerned, the possibility to assign, compare, and condense data in one consistent structure is a must,’ he added.


Maximising quality Looking ahead, Hiltner believes it is not inconceivable that, rather than being used solely for inspection purposes, vision data could one day be used to correct or adapt production processes automatically – perhaps if a particular machine falls outside specified parameters. Although he admitted that adapting a


production process is a very complex task, which normally takes a lot of time as well as highly qualified personnel, Hiltner also predicted that in the future it would be imaginable that self- learning methods could make reliable conclusions out of defect types, helping to adjust the machine settings of each production step automatically. Meanwhile, Chabot was keen to stress that the machine vision systems installed in


32 Imaging and Machine Vision Europe • Yearbook 2018/2019


manufacturing environments 15 years ago were not what they are today. In particular, he highlighted the fact that adoption was not high, repeatability was problematic, and lighting was oſten an ongoing battle. ‘Flash forward to now and we see more and


more inspections being done with vision systems that provide robust results day in and day out,’ he said. In his view, the same perspective also holds


true for vision data, which he observed was not on anyone’s radar back then. However, in the contemporary workplace, he pointed out that vision systems generate a large volume of data, which he argued can and should be leveraged to help improve yield and performance for manufacturing environments. Moreover, although he believes that people remain key in performing most of these analytic tasks to achieve results, he stressed that there is a trend building towards artificial intelligence taking on that role. ‘We know that machine learning starts with


data and the more data you have, the higher the learning potential. Getting data from every part of the manufacturing process – vision data included – is a necessary step towards artificial intelligence becoming commonplace in manufacturing,’ he said. ‘Currently, machine vision data – not the


images, the actual data – still tends to be buried in the black box,’ he added. ‘While there are ways to extract it, the interface varies dramatically from vision system to vision system. Tis is something that will need to be overcome to really shiſt machine vision to the forefront of manufacturing analytics. Once that happens, vision systems can and should be a prominent piece of the process by which manufacturing lines self-adjust to maximise quality and output.’ O


@imveurope


www.imveurope.com


MVTec Software


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