Machine Vision 2021/22 - Machine Vision Key Technology for Automation Solutions Applications - Products - Suppliers - VDMA Verlag
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Machine Vision Key Technology for Automation Solutions Machine Vision 2021/22 Applications – Products – Suppliers
CONTENTS 01 Contents Contents 02 Machine vision – when machines see and comprehend 04 The future of industrial production 06 From product quality to quality of life 13 The right solution for every job 13 Diverse systems 14 The most important trends 16 Standards for machine vision 17 Ensuring your vision project is a success 18 OPC Machine Vision 20 Company profiles 48 Index of members 53 Imprint
02 MACHINE VISION Machine vision – when machines see and comprehend Cameras generate images that are automatically Safe production, reliable products analyzed by software. Machine vision systems perform quality tests, guide machines, control Machine vision guarantees safety – in your production process as well as in the processes, identify components, read codes and end product. deliver valuable data for optimizing production. Machine Vision is constantly capturing new fields of application – also beyond the arena Sustainability of manufacturing – always striving for higher Smooth material flow, considerate use of energy quality, efficiency and product safety. and resources – machine vision makes it possi- ble. At an early stage in the production process faulty parts are detected and sorted out before they are processed further. It prevents prevents 10 reasons to use defective components and a high reject rate of products. machine vision Stable and optimised processes 100% quality check for maximum product quality Recognizing trends and irregularities in pro duction processes early on – machine vision Producing quality non-stop, 24 hours a day, paves the way for realizing the smart factory of 7 days a week – expensive product recalls, the future. product liability claims and image loss can be avoided. High savings potential Higher productivity Machine vision systems lower your costs. They and competitiveness often pay off after only a few months. Modern production is automated. Machine vision enables manufacturing companies to Flexibility remain competitive, to prevent an exodus of key technologies, to generate qualified jobs, Modern vision systems are flexible, able to and to capture new markets. Not just since learn and to adapt. This is ensured by self- Covid-19 has it become clear that production is training and standardised interfaces software. increasingly carried out where consumption “Plug and play” – this is what the machine takes place, not least because of rising environ- vision industry is working on. Even batch size 1 mental and transport costs. Production becomes becomes feasible. competitive through the use of machine vision. The outsourcing of production to low-wage countries can be avoided. User-friendliness Specific programming skills were perhaps previously required. Ease-of-use and seamless integration into the production process have been commonplace for a long time now.
MACHINE VISION 03 Ergonomic workplaces Machine vision – a growth market Checking tasks are often monotonous and tiring The machine vision industry in Germany and for the eyes. Sometimes they are too complex for Europe has been reporting high growth numbers the human brain. Monotonous and dull tasks are and record sales figures for years. Between 2013 accomplished by a „seeing machine“ – machine and 2019, sales in the German industry grew by vision systems support workers, ensure a perfect an average of 9% per year. In 2020, sales declined human-machine interaction resulting in a due to the COVID-19 pandemic. But far from more advanced and safe workplace. being as drastic as in other industries. The reason for the boom: Equipped with machine A technology serving people vision systems, machines and robots are learning to “see”. Within the global race towards greater With and for humans, ensuring security, higher automation, this key technology is not just found quality, and safety in and outside the factories in traditional industrial sectors, but is conquering worldwide. Optimising traffic flows, the perfect brand new areas too. Improved quality, greater swing in golf, training of medical doctors, check- reliability, increased safety and cost-effectiveness ing moles, waste separation and recycling – all are benefits that are just as crucial in non-manu- benefit from machine vision! facturing contexts as in the realm of industrial production. www.seeing-machines.eu Machines with eyes wide open You want to know more about machine vision? Then take a look at the VDMA multi-media report about „seeing machines“.
04 MACHINE VISION The future of industrial production The factory of the future will be designed From inspection to optimization to yield maximum benefit from the lowest possible consumption of material and energy. Many machine vision systems today are more than pure inspection systems, as they make it pos- Interconnected and more flexible production sible to recognize trends in production processes processes will pave the way for new business early on. Thus quality control is evolving into pro- models. All devices will be able to communicate duction optimization, which allows to flexibly directly with each other and thus create true react in the intermediate stages of production, for interoperability. example by implementing warning thresholds in the software for the classification of error types, sizes, dimensions and many other inspection crite- Strict quality specifications are commonplace ria. This allows to detect at a very early stage any today. Random quality checks are increasingly changes that are slowly creeping in. The documen- replaced by a 100% quality control approach. tation of characteristics that are relevant for qual- Each and every stage of production becomes ity, helps uncover the cause of an occurring error. subject to comprehensive documentation and The identified error can then be effectively traceability, especially when the product is safety- avoided. Should errors occur nonetheless, it is critical. This is economically feasible thanks to increasingly possible to implement automated machine vision systems that undertake infallible rework of the faulty workpiece. quality checks along the production cycle. Expen- sive product recalls, product liability claims and image loss can be avoided. The Industrial Internet of Things, Industry 4.0 and Machine Learning Global competition is becoming fiercer: only those production facilities that are among the Machines, production resources and products leaders in terms of productivity and costs will are increasingly directly communicating with be able to survive. Machine vision plays a key role each other and are interconnected; starting with in this because it offers solutions suited to meet the customer’s order to quality control, from the challenges of the future. the shop floor to the board room. The resulting
MACHINE VISION 05 data convergence will open up entirely new Interoperability, connectivity and standards intelligent production options in which all are key factors objects influence and control each other by means of autonomous data and information Interoperability and the ability to interconnect exchange. The “Smart Factory” resulting from are key factors when it comes to implementing this will follow a completely new production Industry 4.0. But standards are crucial: Infor paradigm: Products are identifiable, can be mation networks between companies and localised at any time and know their history, straightforward integration of different value their current state as well as possible options on chains are only possible if all those involved their way to finalisation. C ustomized one-off agree on standards. They define the mechanisms products, manufactured fast and efficiently, of cooperation and the information to be will thus become reality. Moreover, artificial exchanged. Like many players, VDMA supports intelligence, especially machine learning, will the communication architecture OPC Unified further optimize production. Deep learning Architecture. Currently around 30 branch specific algorithms continue to improve the performance OPC UA companion specifications are being of machine vision systems, allow for intuitive developed within VDMA, all of which work in “teaching” of inspection tasks, help find anoma- cooperation with each other in order to ensure lies and enable predictive maintenance. seamless interoperability with each other in the near future. Machine Vision: In September 2019, Part 1 of the OPC UA for key technology for Industry 4.0 Machine Vision Companion Specification (in short OPC Machine Vision) was published (Part 1: Machine vision is the trailblazer and key tech Control, configuration, management, recipe nology for these interconnected production management, result management). Part 2 is processes. No other component is able today currently being developed. The industry agrees to gather and interpret as much data as which information, data, functions and services machine vision systems. The idea is to verify are to be integrated into a production network and process what has been “seen” and transmit and mapped in OPC UA. The aim is an easy the results to the systems of the value chain in connection and straightforward integration every phase of production. This is not only about of machine vision systems into production con- finding out whether a part is good or insuffi- trol and IT s ystems. cient but about triggering intelligent actions. Machine vision is perfectly suited to do this OPC Machine Vision describes the interface because it is the “seeing element”. Increasingly, between machine vision systems and other autonomous guided vehicles (AGV) and autono- systems in an automation scenario. Previously, mous mobile robots (AMR) are roaming in the the system integrator had to develop proprietary factory floor to transport things to the right interfaces for each machine vision application. place at the right time. Machine vision provides The working group has now defined a uniform them with the sensory capabilities to navigate interface that will significantly simplify the intelligently and safely in areas with constant entire integration process in the future. Stand- movement of people. A new concept using many ardised integration of machine vision with all mobile robots is the so-called “matrix produc- its possibilities will advance it from a mere tion”. It consists of robot cells that are connected inspector to a true production optimizer. by mobile robots which bring tools and work- pieces where they are needed – seemingly guided by an invisible hand, but in reality guided by machine vision and AI.
06 MACHINE VISION From product quality to quality of life In industrial production, machine vision An acoustic camera makes sounds visible displays the qualities of a genuine all-rounder. Moreover it has proven itself in many other Integrated into the image processing software, an acoustic camera can be used to determine fields of application beyond the factory. and visualize not only the presence, but also the Improved quality, greater reliability, increased exact location of an error source. Data obtained safety and cost-effectiveness are benefits that in this way can be used to draw precise conclu- are just as crucial in non-manufacturing contexts sions about problems in the manufacturing pro- as in the realm of industrial production. cess or to generate detailed repair instructions. An acoustic camera is used, for example, to check plug connections. In addition to the visual appear- Are all pins in the right place? ance of a plug connection, the noise that occurs during the assembly process and its location can This question is often only clarified when all help to assess whether it has been executed cor- the parts are assembled. But this is often too rectly. In many cases, the mounting position is late and could lead to defective products, for not ideally visible or not visible at all for a stable example. evaluation by purely optical means during the process. Especially in cases like these, the signal With the help of machine vision, it can be offers a variety of additional evaluation options determined in a matter of millisecons whether to ensure the quality of the plugging process. the pins will fit into the PCB’s hole pattern. Source: NeuroCheck GmbH For this purpose, a high-precision image of the circuit board is created. This considerably simpli- fies the testing process for the pin inspection of connector assembly lines, as faulty connectors Absolutely in-line with correlation-free are sorted out before assembly. Good, because metrology who would want a smartphone that does not work reliably? In-line measuring systems in the automotive industry monitor the production process and Source: senswork GmbH provide production plant operators with trends regarding process stability and quality.
MACHINE VISION 07 With high-precision and fast 3D sensors, fea- Surface inspection tures such as bolts, holes and edges can be measured directly in the line in milliseconds. The In industrial production and quality assurance, sensors enable the inspection of the quality-rel- surfaces often have to be inspected, either dur- evant features required for inline process control ing production or in random samples of the fin- in production. As a result of the new tracking ished product. Various 3D sensors, such as laser technology, which records the sensor position on scanners and SfS sensors working according to the robot arm with high accuracy regardless of the “shape-from-shading” principle, are used for temperature influences, metrological tasks can this purpose. Three main areas of application be performed in a cell in the production line in can be distinguished: addition to process monitoring. Thus, time-con- suming correlation measurements with coordi- • Form test: In a 3D image of the surface, nate measuring machines in a dedicated meas- eviations from the shape, outer contours, and d uring room are no longer necessary. This means often also flatness must be checked. that reliable, traceable measurement and test Identification of defects: This is about finding data are available from the first component pro- and evaluating local defects. These can be duced during production start-up. In addition, scratches, bulges, blowholes, burrs, and bursts, the new tracked and traceable in-line measure- etc. ment technology offers the potential to shift Roughness test: Surfaces always show a periodically occurring metrological measure- more or less pronounced roughness. Depend- ment tasks into the line, which contributes both ing on the application, a maximum roughness to a significant acceleration of responsiveness is permissible, or a defined roughness is due to the greatly increased measurement fre- required. quency and to an increase in the efficiency of production processes. This results in a much This technology is used in many applications: higher measuring frequency for quality assur- from surface testing of aluminium, checking the ance and much higher accuracy for process con- roughness of textured lacquer samples during trol. production, to testing Braille characters, or even measuring wrinkles after washing processes. Source: Carl Zeiss Automated Inspection GmbH Source: in-situ GmbH
08 MACHINE VISION Weld visualization using 3D-welding helmet ture of the visitors is checked with the help of a contactless thermometer. Anyone who has an For many years welding helmets have protected appointment receives safety, hygiene and data welders´ eyes at the cost of visibility and produc- protection instructions and has to agree to them tivity. While worker safety is paramount, this before the appointment. The robot informs the reduction in capacity represents one of the big- host about the visit and motivates the visitor to gest limitations to productivity in the field. To disinfect their hands in the waiting area. achieve better results, the Xtreme Dynamic Range (XDR) technology is introduced for creating the Source: Pi4_robotics GmbH revolutionary 3D-welding helmet that enables operators to examine their work in real time. The helmet equipped with XDR acquires and synthe- sizes images as a stereo camera unit. As a result, Lenses enable high-precision adjustment of welders can weld more reliably due to the ability positioning tasks to view weld beads, items being welded and the working environment simultaneously. In the electronic assembly process, or in the space shuttle boom arm: highest precision is Source: XIMEA GmbH needed. The perfect solution to meet high preci- sion requirements of positioning tools are large format lenses with a resolution of up to 100 line pairs per millimetre (lp/mm) and an excellent Concierge robots optical stability over the image field, be it on earth or in space. Concierge robots protect your employees and customers from risks, especially in times of Die-Attach is a lead-free process for bonding COVID-19. They greet guests politely and profes- silicon chips to the pad or die cavity of the sionally. They are reliable and always in a good support structure of microelectromechanical mood. A cloud-based appointment manage- systems (MEMS). Special pick-and-place tools ment system informs the robot which visitors are required to retrieve the die mount from are currently expected. The identity of the visi- the wafer tape and position it on the adhesive. tors is confirmed with facial recognition and a A critical point in production comes when comparison with their identity card data and the adhesive needle must hit a target on the passport photo. In addition, the body tempera- substrate within 0.3 µm. Depending on the
MACHINE VISION 09 machine, the distance from one target to Returning bottles in the supermarket another can be up to 40 mm. The perfect solu- tion to meet the precise requirements of posi- Modern reverse vending systems, which are tioning tools in die-attach machines are large mainly used in food retailing, nowadays work with format lenses with a very high resolution and high-performance camera systems for the recogni- an excellent optical stability over the image tion of barcodes and security features. Within one field. Equipped with a beam splitter, the coaxial second, up to 1,000 images from a total of six illumination option provides uniform illumina- cameras are generated, evaluated and compared tion in all RGB channels without reflections from with a database of more than 35,000 data records. the highly reflective substrate. Lenses without Such a high data processing rate enables an input beam splitter can be used with ring lights for speed of empty containers of up to two metres per incident brightfield illumination. second. Thanks to specially matched LED lighting elements, even very small and low-contrast bar- Source: Jos. Schneider Optische Werke GmbH codes can be reliably detected and evaluated. In addition, the camera-based continuous object tracking system can be used to monitor any posi- tion modifactions or withdrawings of the bottles. Gripping into the box? child’s play! An attempt to defraud by pulling out already deposited bottles or goods – also known as the What is simple for an infant can be a difficult thread trick – can thus be effectively prevented. task for a robot. The complex “reach into the box” with the help of optical sensors is consid- Source: Diebold Nixdorf Technologies GmbH ered a particularly complex challenge in auto- mation. Process stability and high performance, even at Apples or pears? top speeds - this challenge has been worked on Clearly recognized thanks to 3D vision and AI for years. Machine vision technology is not only used for the “ gripping”, but also for optimised Conventional image processing methods some- robot path calculation and the process-safe times face great challenges if they are to reliably multi-pick function. detect variants of an object or product that differ from one another in shape and color. This is Source: ISRA VISION AG often the case when recognizing and sorting
10 MACHINE VISION imaged scene. The shape information replaces the color information of a 2D RGB image, which in turn significantly simplifies the recognition and d ifferentiation of differently colored fruits and enables additional applications, such as the exact positioning and measurement of the recognized objects. Source: BASLER AG Modern approach to farming using depth cameras Depth cameras have proven themselves extremely effective in vertical farming and auto- mated harvesting solutions. The information they provide can be used to monitor plant growth and health. Harvesting machines can better understand their surroundings and navi- gate efficiently. They are able to locate fruit on a plant, assess its ripeness and pick it, all without damaging the plant. Adding a 2D RGB sensor to provide colour information further augments the captured 3D data, e.g. to better assess a plant’s health or the ripeness of the fruit. The visible spectrum and 3D data are used for agri- cultural machinery navigation systems, to avoid plants and other obstacles. Source: FRAMOS GmbH fruits, for example. With the help of trained neu- ral networks and 3D image processing, not only can different types of fruits be precisely recog- Stone collector 4.0 nized and distinguished from one another, but also different varieties of a species, such as green Finding and removing field stones automatically apples from red ones. is possible thanks to image processing. How does it work? First, a drone conducts a survey using advanced A deep learning-based vision system uses 3D sensors, GPS, and trained neural networks using data as raw material. This system delivers high- computer vision to accurately map rocks’ loca- resolution 3D images with nearly millimeter tions and sizes. Then there is a hydraulically precision. A grayscale image as an intensity controlled implement mounted on a front-end image is supplemented by distance measure- loader. A human then drives the tractor on the ments for each individual pixel using time-of- optimized route provided by the map, and the flight measurements of light pulses in the near computer vision on the implement automati- infrared range. The resulting 3D point cloud cally identifies the rock and pulls it into the provides additional information about the tractor bucket.
MACHINE VISION 11 The core components of both systems (drone and tractor) are robust cameras whose high- resolution images are used in three areas: for accurate training of the neural networks, for identification of the stones on the aerial photo- graphs and for real-time recognition of the field stones on the robot arm, which picks them up from the field. Source: LUCID Vision Labs GmbH Sensor fusion and artificial intelligence for the safe mobility of the future How will we travel or transport our goods tomorrow? The answer is as simple as it is complex: autonomous, without drivers or pilots. The machines will decide, not humans. What sounds like a utopia is already in parts reality Recycling – today. no problem thanks to Hyperspectral Imaging In addition to public acceptance, safety will Applications in the recycling industry are becom- be the key to success. This can only be achieved ing more and more demanding. Not only a suit- if various sensors that are intelligently fused able data acquisition and data processing is together support the decision algorithms of required, but also powerful data analysis and artificial intelligence. Whether for cars, delivery data evaluation. robots, air cabs or drones. High-performance industrial cameras are just as important as Color cameras can only process information in lidar lasers or radars. Due to their individual the visible light band. Hyperspectral imaging in capabilities, they guarantee a safe flight, even the near infrared range offers the degree of clas- under varying weather. Vehicles that have sification accuracy required for many recycling limited on-board sensor technology can be tasks by directly measuring the chemical molec- upgraded by customized hardware and software ular composition by means of light absorption solutions, e.g. as a retrofit, and automated safely (spectroscopy). Thus different materials can be and efficiently. This not only enables drones to measured exactly according to their chemical fly autonomously, but also to perform real-time compositions in real time. operations. For example, centimeter-precise 3D maps can be generated, or GPS-free localization Due to a combination of a powerful real-time and exact environment recognition can be embedded vision system in the relevant wave- enabled. Automated inspection and monitoring length range and a real-time analysis software missions are already reality and enable high- equipped with a user-friendly interface, the precision data evaluation in real time today. degree of classification accuracy necessary for such applications can be achieved. Source: Allied Vision Technologies GmbH / Spleenlab GmbH Source: EVK DI Kerschaggl GmbH
12 MACHINE VISION Safe train rides – Frequent malpositions and external influences thanks to image processing are detected and evaluated in order to implement preventive maintenance measures – not only on When a train drives into the maintenance hall, the train but also on the railroad track. camera gates automatically inspect the train. Thereby any damage or deterioration of the roof, The system “learns” to understand correlations side and underfloor structure is being detected. better and better the more data is collected. This Using different imaging methods, it is possible to saves valuable time and ensures that mainte- carry out inspections which the human eye can- nance can be planned precisely. From the selec- not perform. All findings are carefully docu- tion of components to the exact time of ordering mented - within minutes. An analysis tool links and an optimised stockkeeping of spare parts. the recorded data of the maintenance process with other collected information. External param- Source: PSI Technics GmbH eters can also be included, such as weather con- ditions, travelled distances or velocities. Based on Big Data and Machine Learning, this Success in the saddle huge data can be exploited to its full potential. Safety-relevant components and their diagnoses Cycling is a popular sport. Under exertion, how- can be classified together so that correlations ever, it exposes the body to considerable stress. become clear and conclusions can be drawn as to Pain, especially in the back, knees and hips, is when parts need to be replaced. common among competitive athletes. A bike-fitting system with a camera and suitable analysing software helps professional and ama- teur athletes to position themselves optimally on the bike. This prevents incorrect postures, mini- mizes the risks of injuries, prevents pains and maximizes athletic success. For analysis, the bicycle is fixed in a roller trainer. A camera records this set in a two-dimensional X-Y axis. While the cyclist pedals with different intensity and in different positions, the camera records the sequence of movements. Changes in the joint angles are recorded in high quality and at high frequency (60 frames per second), pre- cisely and without any distortion. A motion anal- ysis software measures the direct effects on the respective mechanical parameters of the cyclist, i.e. influence of force, movement and pedalling technique. Source: IDS Imaging Development Systems GmbH
MACHINE VISION 13 The right solution for every job The VDMA Sourcing Service helps you Choose between three comfortable search identifying the right partner for your machine strategies: vision application – no matter whether you need components, complete systems • Search by type of product (systems, components, services) or services. • Search by type of application (e.g. surface inspection, robot vision, identification) • Search by type of customer industry www.vdma.org/visionfinder Diverse systems Some machine vision systems are so compact you can hold them in the palm of your hand. Sometimes, they are integrated vision systems that work intelligently directly from devices and enable them to see and understand. Others could fill a whole room. • Application-specific vision systems are turnkey • Smart cameras (intelligent cameras) are systems for a specific application such as c omplete vision systems that include software the inspection of flat glass or wafers. They are for image analysis, all contained within the usually PC-based and offer high-performance. compact housing of the camera itself. They can be programmed flexibly to carry out different • Configurable vision systems are usually types of tasks. C-based as well. Unlike application-specific P systems they are more versatile in their use. • Vision sensors are also complete vision systems End users can often implement a range of within a single compact housing. In contrast to applications via a graphical user interface. smart cameras, they are made for one specific type of application, e. g. code reading. System integrators, with their high level of sector- and application-specific know- Making use of versatile components, solutions can quickly be developed to how, offer made-to-measure solutions. tackle a wide range of tasks.
14 MACHINE VISION Reaching into the box - child‘s play! Robot Vision What is simple for an infant can be a difficult task for a robot. But with machine Machine vision delivers sharp eyes for industrial robots and teaches them vision, the „reach into the box“ works! to do see and understand. Robot Vision is an important trend, also in robotics. 3D technology is being used here. The most important trends Machine vision is developing rapidly, creating Performance ever greater benefits for users. The efficiency of machine vision systems is rapidly increasing, thanks to higher-definition cameras, steadily increasing processor p erformance as well The third dimension as multi-core processor technology, state-of-the- art software and standardised interfaces. Ever With machine vision becoming three-dimen- more rapid inspection speeds and accuracies are sional, many tasks can be accomplished in a thus achieved at comparable costs. cost-effective way – from the precise inspection of adhesive beading or welded seams and auto- mated gripping of unsorted items in boxes Colour through to non-contact precision measuring of formed sheet metal parts in the production In many applications, colour recognition is of cycle. great advantage. For example, different models or components can be identified or sorted, and quality checks can be conducted on the basis of Standardisation colour. Colour recognition has now become a standard routine for which numerous machine Standardised interfaces facilitate the integration vision solutions are available. of individual machine vision components into an overall workable system. This reduces the amount of effort required and makes the machine vision solution efficient and inexpensive.
MACHINE VISION 15 User-friendliness Embedded Vision Whereas in the past experts were needed to Embedded Vision enables image processing on implement machine vision solutions, today many compact, high-performance, and low-energy adaptations can be carried out without any great computing platforms. They can be integrated prior knowledge, thanks to intuitive configura- in areas of limited space or small devices. This tion options and ergonomic software user inter- technology thus opens up many new fields of faces. application which so far could be covered neither by PC-based nor by intelligent machine vision systems. Hyperspectral Imaging Hyperspectral cameras or sensors take several Machine Learning images of a scene in different wavelength ranges. Combined, the images provide a greater Machine learning is an important area of artifi- depth of information. This technology is used in cial intelligence. Computer programs based on areas and applications whenever ingredients or machine learning use algorithms to find solu- substances need to be identified and discerned, tions to new problems. The artificial system rec- and are not recognizable through a standard ognizes patterns in the learning data it receives. color or monochrome picture. For instance, in the Deep Learning enables the classification of food or wood industry, in recycling, mining or images with a better classification rate than with agriculture. previous methods. The user does not have to explicitly program the different error classes, i.e. you just “show” the deep-learning algorithm images and it learns from them. Would you have seen the scratch? Modern vision systems detect errors that the human eye tends to overlook or classify as irrelevant. As the example of the saxophone shows, such faults would be per- ceived as annoying when touching or playing the scratched key.
16 MACHINE VISION Standards for machine vision Standards facilitate use of machine vision and integration of individual components into a complete system with optimum functionality. They cut down on development time and investment costs and also accelerate time-to-market. And, the number of standards is growing steadily. You will find an overview of existing globally accepted machine vision interface standards in the brochure “Guide to Understanding Machine Vision Standards” which can be downloaded under www.vdma.org/vision About G3 – Global Coordination of Machine Twice a year, technical experts meet in so called Vision Standardization „International Vision Standards Meetings“ for joint work and coordination according to a rotat- In 2009, member supported trade associations ing system alternating between Asia, America from Asia, Europe and the USA began a coopera- and Europe. Existing G3 standards are: GenICam, tive initiative to coordinate the development of EMVA 1288, GigE Vision, CoaXPress, Camera Link, vision standards and founded the so called “G3” Camera Link HS, USB3 Vision and VDI/VDE/VDMA initiative. The aim of G3 is to develop globally 2632 and OPC Machine Vision. The newest adopted vision standards and hence to avoid G3-standardisation projects are emVision and the overlapping or duplication of efforts. Open Optics Camera Interface.
MACHINE VISION 17 Ensuring your vision project is a success Part 3: Acceptance test of classifying machine vision systems For measuring (non-classifying) machine vision systems, quantitative capability analysis is already well established. Measurement uncer- tainty is usually employed as an indicator. Until now, on the other hand, there have been no corresponding and accepted qualification indicators for classifying machine-vision systems whose results are attributive variables. Part 3 closes this gap and introduces indicators describ- ing the classification capability of a machine www.vdma.com/flawless – vision system. The preparation of a requirement and system specification ensures successful machine vision applications. Part 4: Surface inspection systems in flat steel production The VDI/VDE/VDMA 2632 series of standards structures the communication between supplier This standard describes concepts for continuous and user. The standards help to avoid misunder- performance monitoring of surface inspection systems. It is intended for manufacturers and standings and to handle projects efficiently and users of surface inspection systems and describes successfully. procedures for stability testing. Although this work is oriented towards the requirements of flat steel production, it is expected that the results Part 1: Basics, terms, and definitions can also be applied to other inspection tasks such as aluminium, paper or foil. Knowing what you are talking about is the start of every successful project. The guideline Publisher of VDI/VDE/VDMA 2632 series is the describes the principles and defines the terms VDI/VDE Society Measurement and Automatic necessary for the use of image processing Control. The standards were elaborated in coop- systems. It defines a consistent terminology for eration with VDMA Machine Vision. The docu- all cooperation partners. ments are practice-oriented, thorough, and spe- cifically address the requirements of machine vision. They include both a German/English and Part 2: Guideline for the preparation of a for part 2 an English/Chinese version, and can requirement specification and a system therefore be applied in the communication with specification customers worldwide. The standard provides assistance in the prepa The VDI/VDE/VDMA 2632 standards can be ration of specifications for industrial image purchased on: processing systems. Particular emphasis was www.vdi.de/2632 placed on the representation and description of influencing factors as well as on their effects. The p roject partners are thus able to identify influences at an early stage during planning and to find optimised solutions. www.vdma.com/flawless
• In parallel to existing automation standards tromagnetic waves Electromagnetic wavesvision Machine Machine Resultvision Result (e.g. measurements, (e.g. measurements, codes codes A vision isible light, IR,(e.g. X-rays) systems visible system light, by the OPC IR, X-rays) Foundation. system and characters, and characters, poses, …) poses, …) 18 MACHINE VISION Vision services Part 2 … n (Vision skills layer) Working Group OPC Machine Vision Working Group Machine Vision Basic services (Infrastructure layer) The “Eye” of Industrie 4.0 The “Eye” of Industrie 4.0 Part 1 Machine Vision Unmatched product Unmatched•product and Machine The and process OPC UAVision process data for Quality assessment and OPC Companion data UA for Specification Vision (in short OPC Machine Vision) provides a generic The scope is not only to complement or substi- tute existing interfaces between a vision system Trigger Status Recipe Resu • Quality assessment • Track & Trace Recipe information model for all vision systems – from and its process environment by using OPC UA, • Track & Trace OPC Vision, part 1 Electromagnetic waves Machine vision Result (e.g. measurements, codes • Process control simple vision sensors to complex inspection (e.g. visible light, IR, X-rays) system systems. Put simply, it defines the essence of and characters, poses, …) but rather to create non existing horizontal and vertical integration abilities to communicate • • Process control Process optimization Partdata1 describes an abstraction process parti of the g 07.06.18 11:21 Machine Vision The scope is not and only to OPC UA or substitute any vision system that does not necessarily have complement relevant to other authorized • Process optimization • Data analytics to be a “machine” vision system. OPC Machine system, i.e. the representation cipants, e.g. right up to the IT enterprise of the so level. It existing interfaces between a vision system Vision is the accepted and officially supported and its process is possible to have a gradual phase-in of OPC • Data analytics environment by using OPC OPC UA, but rather UA Companion Specification to create twin" of the system. It handles the man for visionnon- Machine Vision with coexisting other interfaces. The “Eye” of OPC Industrie 4.0 The configurations benefits are a shorterand time results to marketin by aa standard existing horizontal andsystems vertical Available via OPC UA by the Foundation. integration abilities to communicate • OPC UA For relevant data Unmatched product and process data for to other authorized process INDUSTRIE the4.0 simplifiedcontents integration,stay a generic Working Group vendor-specific applicability and and a Available via the entire factory • Quality and assessment beyond boxes (1). It allows the control scalability STANDARDS and an improved customer perception of a visio • participants, • e.g. right Directly, up•to without For the entire factory and• Process • OPC TrackVision& Trace thecontrol conversion Process beyond IT enterprise layerslevel. It is Machine Vision due to defined and consistent semantics. Specific generalized way, abstracting the necess possible toIn • have a gradual parallel to existing phase-in of OPC optimization automation Visionstandards with example: OPC Machine Vision enables Machine • coexisting Directly, without conversion layers • The DataOPC UA analytics Companion Specification Vision (in short state The basicmachine Vision concept of OPCconcept to speak to the whole factory(2). Vision is a subdivision into several and beyond. other interfaces.OPCThe benefits Vision) provides a are genericainformation shortermodel for all parts. Part 1 includes the basis specification and describes an • Intime parallel to existing to market Machinevision systems - from simple vision sensors to complex automation by a simplified Available Vision and integration, via inspection OPC systems.UAOPC UAit defines the essence of standards a generic Put simply, HMI (SCADA) infrastructure layer which provides basic services in a generic way. From part 2, a vision skill layer is addressed which provi- • any vision For the system entire that does factory not necessarily have to be a and beyond des more specific vision services. applicability and scalability andwithout • "machine" Directly, an improved vision system. customer OPClayers conversion Vision is the accepted and Fundamentals Configuration data Vision system Recipe perception due to defined• vision and consistent officially by the OPCsemantics. supported OPC UA Companion Specification for black box OPC Classic In parallel to existing automation standards Electromagnetic waves Machine vision Result (e.g. measurements, codes Any production IT system systems Foundation. A vision system is any system that has the (e.g. visible light, IR, X-rays)Robotsystem and characters, poses, …) Specific example: OPC Vision enables Machine Vision to capability to Data record and1process digital images The speak “Eye” of Industrie 4.0 Camera IPC EtherCAT management to the whole factory and beyond. Machine vision system or video streams, typically with the aim of Cloud Part 2 … n Vision services (Vision skills layer) Recipe data from this data. The out- extracting information Unmatched product and process data ERP for black box Interface 4 put of a vision system canBasic beservices any image-based king•Group Quality assessment Working SCADA Group ERP MES Fundamentals PROFINET Interface 3 information like measurements, inspectionBehavior (Infrastructure layer) Observe Working Group •ATrack & Trace The “Eye” of Industrie 4.0 Interface 2 Control Part 1 results, processTrigger control data, robot guidance chine•Vision vision system is any system Machinethat has Vision the capability to record PLC Status Recipe Result … 2 Interface 1 Cloud Device data, etc. Process and process control digital Machine vision images application PLCUnmatched product and process data for using • or video Quality streams, typically assessment Automation OPCwith UA gradual •the Process optimization phase-in Machine Vision established interfaces and OPC UA pyramid communication • Track & Trace The OPC basicVision, part of1OPC Machine Vision is a aim of extracting information from this data. The output • Process control concept •ofData Part 1 describes an abstraction of the generic vision Saule_MachineVison_Automatica2018.indd 1 07.06.18 11:21 e Vision and OPC UA analytics a vision system can be any Machine Vision is not and image-based only toOPC UA or substitute optimization information like The scope • Process complement subdivision system, i.e.into several parts. the representation Part of the so 1called includes "digital The possibility of a gradual phase-in of OPC Machine measurements, inspection results, process control data, robot Saule_MachineVison_Automatica2018.indd 1 existing • Data Vision interfaces between a vision system and its processthe basis s pecification and describes an infra- analytics environment by using OPC UA, but rather to create non- twin" of the system. It handles the management of recipes, configurations and results in a standardized way, whereas 07.06.18 11:21 Future parts existing horizontal and vertical integration abilities to structure layer which provides basic services Recipe guidance data, etc. In future the contents stayparts, 2,the vendor-specific generic and are treatedbasic as black inform aves Available via OPC UA Machine vision Result (e.g. measurements, codes communicate Available via OPC UA data to other authorized process in a boxes relevant generic (1). way. It allows In the part the control of vision a vision system system in a shift toway, a more specific "skill-based” inf system -rays) and characters, poses, …) participants, e.g. right up to the IT enterprise level. It is described in part 1 has been enhanced with • Forpossible the entire factory and beyond generalized abstracting the necessary behavior via a • For the entire factory and beyond • Directly, to have a gradual phase-in of OPC Vision with state machine concept (2). mation model. Vision skills could inclu 07.06.18 11:21 without conversionThelayers information pertaining to asset management Machine Vision and OPC UA coexisting other interfaces. benefits are a shorter • Directly, without conversion • Intime Recipe tolayers parallel to existing market automation by a simplified standards integration, a generic and condition monitoring. It provides identifica- applicability and scalability and an improved customer tionpresence data detection, completeness inspe Vision system • In Electromagnetic parallel towaves existing automation perception Vision system due Resultstandards to defined and consistent semantics. (e.g. measurements, codes the and Configuration tion, condition components pose black box monitoring parameters for detection, that build up a etc. machine For 1 visionthis purpo Specific example: OPC Vision enables Machine Vision to Data 1 Result data (e.g. visible light, IR, X-rays) The speak “Eye” of whole Industrie 4.0 and characters, poses, …) management black box e” of Industrie 4.0 HMI to the (SCADA) factory and beyond. the proprietary system. Recipe data input and output data b Unmatched product and process data for black box tched product and process data for lity assessment •Fundamentals Quality assessment boxes will be broken down and substitu Behavior Observe Working Group •ATrack vision&system Trace is any system that has the capability to record Control k & Trace OPC Classic •and Process control process digital images or video streams, typically with with standardized2 information structur ess control •the ess optimization Robot Machine Process aim ofoptimization •ofData Vision extracting information from this data. The output analytics a vision system can be any image-based information like and semantics. a analytics Camera IPC EtherCAT Future parts measurements, inspection results, process control data, robot Machine vision system guidance data,OPC etc. UA In future parts, the generic basic information model will Available Cloud via le via OPC UA Basics of Machine Vision • For the entire factory and beyond ERP shift to a more specific "skill-based” infor- mation model. Vision skills could include Saule_MachineVison_Automatica2018.indd 1 07.06.18 11:21 18 Machine Vision and OPC UA Interface 4 the entire ERP factory and beyond • Directly, MES without conversion Recipe layers More Info at ctly, without conversion layers Interface 3 presence detection, completeness inspec- https://bit.ly/2CM5ev6 • In Electromagnetic PROFINET SCADA parallel towaves existing automation Vision system standards Result (e.g. measurements, codes tion, pose detection, etc. For this purpose, arallel to existing automation standards Interface 2 (e.g. visible light, IR, X-rays) and characters, poses, …) Interface 1 PLC HMI the proprietary input and output data black Cloud (SCADA) PLC Device boxes will be broken down and substituted Machine vision application using Automation OPC Classic gradual OPC UA with standardized information structures phase-in Robot and semantics. The “Eye” of Industrie 4.0 established interfaces and OPC UA pyramide Camera IPC communication EtherCAT Machine vision system Cloud ERP Unmatched product and process 18 data for Interface 4 ERP MES PROFINET Interface 3 SCADA 07.06.18 11:21 Interface 2
MACHINE VISION 19 DUSTRIE 4.0 TANDARDS OPC Machine Vision, part 1 OPC Machine Vision, part 2 ort The basic concept of OPCPart Vision is a subdivision 1 describes an abstraction into of theseveral generic Part 2 pertains to the Asset Management and or all parts. Part 1 includes thevision basissystem, specification and describes i.e. the representation anso of the Condition Monitoring use cases. For the asset ex infrastructure layer which provides basic services in a generic called “digital twin” of the system. It handles management use case, one of the most impor- tant requirements is to be able to successfully of way. From part 2, a visiontheskill management of recipes, configurations and layer is addressed which provi- results in a standardized way, whereas the con- identify the asset itself. The components that ea des more specific vision services. tents stay vendor-specific and are treated as black build up a machine vision system have also been and boxes (1). It allows the control of a vision system considered as assets in this case. The condition on for in a generalized way, abstracting the necessary monitoring use case forms a basis for business behavior via a state Any production IT system machine concept (2). cases like predictive and preventive maintenance. Therefore, part 2 also defines the parameters for the identified components of a machine vision system, monitoring and analysing which can Part 2 … n Vision services provide us with useful information needed for (Vision skills layer) maintenance and diagnostics of the system and its components. Basic services Any production IT System (Infrastructure layer) Future parts Part 1 Trigger Extended Status Vision Services Recipe Result … The aim for future parts is to break down the Part 3..n proprietary input and output data black boxes defined in Part 1 and to substitute them with Basic services Vision system components (focus on methodology) (identification, condition monitoring) standardized information structures and seman- OPC Vision, part 1 tics. Currently a possibility to provide standard- Part 1 describes an abstraction of the generic vision ized structures for the Result data black box is Configu- Identification Recipes Results Topology Components ration & Monitoring being discussed in the working group. system, i.e. the representation of the so called "digital ocess twin" of the system. … It handles the management … of recipes, non- configurations and results in a standardized way, whereas o Part 1 (released) Part 2 (in progress) the contents stay vendor-specific and are treated as black ess boxes (1). It allows the control of a vision system in a t is Conceptual model for OPC Machine Vision generalized way, abstracting the necessary behavior via a ith state machine concept (2). er Configuration data Vision system black box 1 on to Data 1 Result data management black box Recipe data black box Behavior Observe Control cord 2 h tput Main scopes of OPC Machine Vision, part 1 ke obot Future parts In future parts, the generic basic information model will
20 MACHINE VISION Company profiles Company profiles 21 Allied Vision Technologies GmbH 22 ASENTICS GmbH & Co. KG 24 Basler AG 26 Baumer GmbH / Baumer Inspection GmbH 28 Balluff GmbH 29 DVC Machinevision b.v. 30 EngRoTec-Solutions GmbH 31 i nos Automationssoftware GmbH 32 I DS Imaging Development Systems GmbH 34 ISRA VISION AG 36 iim AG 37 MATRIX VISION GmbH 38 LUCID Vision Labs GmbH 40 MVTec Software GmbH 41 NeuroCheck GmbH 42 OCTUM GmbH 43 Polytec GmbH 44 Vision Components GmbH 46 SVS-Vistek GmbH 47 VMT Vision Machine Technic Bildverarbeitungssysteme GmbH
Allied Vision 21 Focus on what counts – digital industrial cameras for your application For more than 30 years, Allied Vision has been We know how to help you find the best helping people to reach their imaging goals. We camera solution for your application. That don’t just develop cameras. We provide answers. includes a digital camera, but also the right lens, the right connectivity hardware and the right software interface. Our job is to reliably deliver Allied Vision supplies camera technology and the image you need, when you need it and how image capture solutions for factory automation, you need it. industry and science, traffic monitoring and many more application areas in digital imaging. To deliver the precision and reliability our custom- By focusing on what counts for each customer, ers need, we manufacture our products according Allied Vision finds solutions for every application, to the highest quality standards in state-of-the- a practice which has made Allied Vision one of art clean room facilities. All our manufacturing the leading camera manufacturers worldwide in sites are ISO 9001 certified and comply with the the machine vision market. ISO 13485 quality standard for medical devices. Our three-year warranty reflects our commitment Our engineers design digital cameras with a large to quality. scope of resolutions, frame rates, bandwidths, interfaces, spectral sensitivities, sensor technolo- Our worldwide sales and support network allows gies, and technical platforms. We have created a us to deliver first-class service before, during and modular concept to ensure that your camera after your purchase. We have offices and sales adapts to requirements of your application and representatives in Germany, the UK, France, the not the other way around. USA, Singapore, and China and have teamed up with selected distribution partners in more than Allied Vision is at your side throughout the life 40 countries to ensure we are always there to cycle of your image-processing project. help you whenever you need us. Allied Vision Technologies GmbH • Allied Vision is a member of the TKH Group Taschenweg 2a • 07646 Stadtroda • Germany Phone +49 36428 677-230 • E-Mail info@alliedvision.com Internet www.alliedvision.com
22 ASENTICS Providing solutions that ensure quality – consistently, effectively and economically ASENTICS – 360° inspection and 100% quality control using standardized OPC UA Interface (VDMA 40100-1) for Industrie 4.0 communication By definition, quality is the match between ASENTICS is one of the leading application- the expectations of a product and its properties. oriented image processing companies in Europe. Producers in all markets are facing major We have been providing solutions for optical/ visual quality assurance in all fields of production challenges because of our globalised world, for more than 20 years, from small, intelligent the desire for 100% quality is continuously stand-alone devices for simple tasks right facing the demand for competitive prices. To through to universally applicable modular and accomplish these tasks, ASENTICS has created scalable image processing systems for in-line solutions that grow together with the monitoring. demands and meet the highest standards ASENTICS is excellently prepared for the of the quality checks. challenges of Industry 4.0. We consider machine vision as not only a forerunner, but also a key technology. The images have to be verified at every stage of production, processed and made available to the systems operating in the value network. Apart from making a statement on a good or bad part, ASENTICS machine vision products also manage intelligent actions down the line.
You can also read