The VISION show, the machine vision industry’s bash which takes place every two years near the charming City of Stuttgart, rarely disappoints and this year’s event was a cracker.
On my rounds of the show floor, I visited a large number of companies, some of which I include here in our feature.
If you mention software in the machine vision sector, one company on everyone’s lips is MVTec, but is the pressure of staying ahead in a competitive sector, getting to them? Not so, said Dr Olaf Munkelt, Managing Director, who was in a buoyant mood about MVTec’s prospects and the sector in general.
MVTec used VISION to present its entire range of cutting-edge technologies, including the latest iterations of HALCON 18.11 and MERLIC 4 software. It also showcased future-oriented topics such as deep learning, embedded vision, deflectometry, process integration, 3D vision and hyperspectral imaging.
HALCON 18.11 was out to win support with what the company described as numerous improvements and new functions, some of which are based on deep learning, an artificial intelligence methodology. The software is now out and offers expanded embedded vision capabilities. The release also provides helpful new functions for developers.
On the booth there was a live demo with an UR3e robot, demonstrating an industrial application scenario with HALCON 18.11. The robot arm reached into a collection of objects and was able to locate the position of the relevant object, thanks to HALCON’s matching technologies. The arm precisely removed the object from the crate, recognizing it by means of innovative deep learning functions, and then sorted it.
Another demo showed HALCON’s compatibility with many important hardware standards: three embedded boards and a smart camera were set a machine vision task. The corresponding cameras were aimed at an analysis object, which was moved horizontally at high speed. This illustrated how quickly and robustly the standard machine vision software works on a wide range of common embedded hardware components.
There was also a demo of the ready-to-use HPeek system image for the Raspberry Pi. HPeek is MVTec’s license-free benchmarking demo program, which is available free of charge and can be used to evaluate HALCON’s performance on Arm-based embedded platforms in real-life applications.
The inspection of objects with reflective surfaces with the aid of deflectometry was also demonstrated. With this method, reflections of known patterns on the particular object surface were evaluated to locate defects. The demo automatically checks different defective objects in two inspection stations and unequivocally localizes the defects with MVTec HALCON. This shows how HALCON can be used to straightforwardly create a complete system for complex inspection tasks – without any special hardware.
MERLIC 4, the new release of the software for rapidly creating complete machine vision solutions, offers a wide range of functions that make machine vision even easier. MVTec said the software has an outstanding new feature: parallelization, i.e., the ability to run separate tools at the same time.
A live demo documented the practicality of this new function by performing different inspection tasks with the aid of two cameras, demonstrating the parallelization feature. It also showed how MVTec MERLIC precisely recognized different fonts on packaging in just fractions of a second, using deeplearning-based OCR technologies. It also illustrated the seamless integration of a programmable logic controller (PLC) into vision systems using MERLIC. The advantageous integration of MERLIC via Hilscher cards was also demonstrated.
A touchscreen demo presented additional machine vision technologies on all aspects of MVTec HALCON and MERLIC, such as hyperspectral imaging and 3D vision.
VISION 2018 also marked the tenth anniversary of MVTec’s “Image Acquisition Partner” program, which is the largest of its kind worldwide.
I caught up with Mark Williamson, Managing Director (UK) and Group Marketing Director, who was in an upbeat mood on a packed booth.
At VISION, they showcased state-of-the-art demo systems based on their Common Vision Blox programming library. This a powerful programming library for fast and reliable development and implementation of vision solutions, which has been deployed successfully throughout the world in more than 40,000 imaging applications in various industries.
‘STEMMER City’ also offered visitors lots of technical highlights from their partners including
Adlink, Automation Technologie, Components Express, CCS, Gardasoft, JAI, Metaphase, Midwest Optical, Mitsubishi Electric, Perception Park, Smart Vision Lights, Xenics and Z-Laser.
The two labs, which showcased current and future developments, proved to be particular crowdpullers, said STEMMER IMAGING. The Embedded Lab focused on future possibilities in the field of embedded vision, whilst the Future Lab showcased solutions on topics such as hyperspectral imaging and camera-based assistance systems.
Sony Europe’s Image Sensing Solutions division had on show its XCL-SG1240 series of Camera Link V2.0 modules, which utilise a 1.1-type global shutter CMOS sensor to deliver 12 MP images at 20 fps.
The series, which includes colour and black/white modules, adds to Sony’s award-winning line-up of GS CMOS machine vision modules. Sony has integrated several features including area gain, shading correction, burst trigger and PoCL functions.
The cameras are suited to sectors that require exceptional detail even in low (0.5 lx), imperfect or changing lighting conditions.
Sony has highlighted intelligent transport (ITS), factory automation, electronics manufacturing as well as AOI/panel inspection as potential applications.
The company has also developed the modules to enable simple migration from CCD systems, allowing for over 1 inch C-mount lens mounts, incorporating the same command control as the XCL-S900 and a Camera Link base configuration.
The modules come with both hardware- and software-triggered synchronisation, with trigger modes of burst, edge detection and pulse width detection.
The XCL-SG1240 series uses the Sony 1.1-type Pregius GS CMOS sensor. The module has been created by Sony’s engineering team to extract the industry’s best image from this market leading sensor.
Arnaud Destruels, European Marketing Manager, Sony Image Sensing Solutions said: “The new module follows months of Sony ISS engineers working with the Sony Pregius 1.1-type sensor and developing a custom module design that extracts the best possible image quality for the machine vision sector.”
A fascinating piece of kit was unveiled by Spain’s Instituto Tecnológico de Informática, which is based in Valencia.
It was the first public demo of what is described as industry’s most versatile machine vision inspection system.
The Scalable Zero Gravity 3D system captures images from all angles and has no blind spots.
The system launches an object vertically into an imaging chamber to precisely capture it from multiple angles at the top of its flight.
This process allows manufacturers to run multiple types of components for analysis in a single batch, and to easily switch components being captured without mechanical reconfigurations.
The technique has successfully completed proof-of-concept testing and is now undergoing commercialisation for a 2019 market launch.
The proof of concept technology has been tested running at 50 parts per minute with a single linear actuator, or 80 parts per minute in a dual-actuator imaging capture device change.
Applications include three-dimensional surface reconstruction with textural analysis, as well as surface defect detection, be it a scratch, stain, crack, corrosion, or geometric alteration.
The system is fully scalable to suit the size of components being measured. The ITI’s prototype on display at VISION used 16 Sony XCG-CG510C GigE modules to achieve this. To cater for larger or smaller components, ITI is able to scale the imaging chamber size to fit special needs.
Cameras are synchronised via the IEEE1588 precision time protocol, with firing timed to match the top of the object’s flight. Also synchronised via IEEE1588 is the imaging chamber’s LED lighting.
Objects enter the imaging chamber from a conveyor belt, which places the component under a polyhedral structure. This is connected to a linear motor which fires the object vertically. Once captured and in free-fall, the object is caught by the same linear actuator moving to match the object’s speed and prevent impact and off-loaded to a second conveyor belt, or tray.
Sergio Navarro, head of advanced industrial vision systems at the ITI, said: “Most machine vision systems inspect a part’s surface only, with the component being held by a manipulator or mounted on a guidance system. This limits the analysis to only a particular segment of the 3D geometry or requires mechanical changes for each new batch, which must also be free of mixtures of different parts,” said “The Zero Gravity 3D not only solves these limitations, it offers a versatility as yet unknown in the industrial inspection area.”
I met Dale Mitt on the Xilinx booth and was given a presentation focussed on the idea of Building the Adaptable Intelligent World, which covered the wide range of the group’s products.
The focus was on how artificial intelligence is driving machine vision markets and how the Xilinx portfolio is geared around these objectives.
Prophesee (formerly Chronocam), inventor of some of the most advanced neuromorphic vision systems, unveiled new capabilities for improving productivity and equipment monitoring in industrial applications. Enabled by its patented, event-based machine vision technology, the company, which was a nominee for the VISION Award, demonstrated advancements in critical machine vision use cases. These included ultra high-speed counting, real time tracking, man machine teaming and vibration monitoring for predictive maintenance, setting a new standard for efficiency and reliability in industrial automation.
The Prophesee booth featured the latest capabilities of its recently introduced Onboard reference system. The system provides a guide for developers to optimally implement Prophesee’s neuromorphic vision technology for a variety of applications. The company has released performance and accuracy specifications for new use cases, including area monitoring, high-speed counting, vibration measurement, or real-time arc welding tracking. Systems that utilize Prophesee’s event-based vision save significant amounts of computational power, bandwidth, memory and energy.
The company’s solutions are aimed at improving a wide range of industrial uses, including accelerating quality assessment on production lines; improving throughput, yield and overall productivity; positioning, sensing and movement guidance for robots to enable better human collaboration; and equipment monitoring (e.g. caused by vibration, kinematic deviations) for predictive maintenance and reduced machine downtime.
Prophesee’s sensors and camera systems open vast new potential in areas such as autonomous vehicles, industrial automation, IoT, security and surveillance, and AR/VR. Its solutions improve equipment efficiency, reliability efficiency and user experiences across a broad range of use models.
It was founded within iBionext Start-up Studio (Paris) in 2014 by Ryad Benosman, Bernard Gilly, Christoph Posch and Luca Verre. The quartet brings a strong combination of experience in image sensing, neuromorphic computing, VLSI design, entrepreneurship and business development.
Prophesee is based in Passage de l’innovation in Paris, with local offices in China, Japan and USA, and has a team of 75 visionary and global engineers. It holds more than 50 international patents and is backed by leading international investors including 360 Capital Partners, Supernova Invest, iBionext, Intel Capital, Renault Group, and Robert Bosch Venture Capital.
Lakesight, now part of TKN, was showing at Vision for the first time.
The company introduces its pitch with the line: “From sensor to solution – the full-service provider of smart machine vision solutions with its own products and development teams presents its portfolio.”
It goes onto to explain: “In industrial image processing, users are faced with a variety of specialized component manufacturers, distributors, system integrators and system builders. Therefore, a variety of partners are usually required to develop appropriate solutions.
“Customers who want to benefit from the enormous possibilities of modern image processing technologies must all too often invest a lot of time, expertise, coordination, and money.
With the breadth of in-house products, systems, technology expertise and in-house development teams, Lakesight is the first vendor in the market to serve its customers at all levels, from the design-in of any image sensor to developing sophisticated, all-in-one artificial intelligence solutions.
“This is made possible by the merger of Tattile, Mikrotron and Chromasens to form the Lakesight group. Therefore, both, small device manufacturers, as well as system integrators, system builders, and major international companies have a financially strong and reliable partner with Lakesight, who understands them and can offer appropriate, technically innovative solutions.”
At the Lakesight booth, its subsidiaries presented their innovative strength with numerous new products in the fields of cameras, embedded vision, lighting and vision PCs.
The specialist for line scan camera technology, Chromasens in Constance, presented the allPIXA evo, the first line scan camera with Dual-10GigE interface. Depending on the model, it delivers 10k or 15k horizontal resolution at 4 lines and is available in a colour, or monochrome version.
With a line frequency of up to 146 kHz, it enables extremely high web speeds and blends seamlessly into the infrastructure of production facilities thanks to the SFP + interface for 10GigE copper, or fiber optic cables.
The new chromaPIXA line scan camera features and integrated module for color conversion into standardized colour spaces such as sRGB or CIELAB in real-time making it ideal for print inspection. For reliable color calibration the software package chromaCalc is included.
The 3DPIXA compact C01-010 is the first stereo camera by Chromasens with a CMOS color sensor, offering an optical resolution of 10 μm, a sample width of 56 mm and a depth resolution of 3.2 μm. Thus, all advantages of the CMOS sensor technology are also available for stereo imaging.
For particularly fast web speeds and short exposure times, high-performance cameras require extremely bright lighting. This is what the Corona series is known for in the market. At the show, Chromasens showed coaxial modules for the Corona series for bright and dark field illumination, as well as two new LED wavelengths in the UV range.
The experts for high-speed cameras from Mikrotron in Unterschleißheim presented four new cameras. Based on the new and faster CoaXPress V2.0 interface, the EoSens 1.1CXP2 delivers up to 3,600 fps at a 1.1-megapixel resolution. The almost 30 Gbit/s are reliably transmitted over 4 channels with 12 Gbit/s each. Compatibility with some new frame grabbers for CoaXPress 2.0 is already verified.
To support customers and applications that previously worked with CoaXPress V1.1, this camera is also available as CoaXPress V1.1 version, EoSens 1.1CXP, with reduced performance accordingly. As an extension to the already available CoaXPress variants, two additional cameras, the EoSens 3FIBER and EoSens 4FIBER provide 3- and 4-megapixel resolution and were equipped with a new fiber optical interface, which allows fast transmission over up to 300m and offer a cost-effective solution for applications with high interference potential or longer transmission distances.
In the Slow Motion area, the portfolio has been expanded to include a portable and immediately operational long-term recorder LTR1,3CL for high-speed video recording at 1.3 megapixels, 506 fps and up to 140 min. extended recording time. It is ideally suited for locating and analyzing quality problems in manufacturing, as well as in research and product development.
Tattile, the Italian specialist for Embedded Vision Systems, introduced two applications which benefit from the low power consumption and processing speed of FPGAs and their seamless integration with powerful CPU architectures. The S12MP Smart Camera features a 12-megapixel CMOS Global Shutter sensor from CMOSIS and an FPGA with 125k Logic Elements where 80% are free for application specific algorithms.
This architecture allows to acquire and process images at 300 fps at full resolution with low latency. Moreover, the FPGA can be programmed via VHDL and the user may implement its proprietary algorithms directly on the FPGA to decrease the load of the CPU. The Linux OS and a dual-core ARM Cortex A9 CPU enable on-camera processing with common third-party software libraries.
The M180 is the successor of the popular M100 GigE Vision Controller based on an Intel core i7 6th generation CPU and Windows Embedded. Six independent GigE ports with PoE allow for multiple cameras connections with maximized performance while the real-time I/Os and encoder input are directly linked to the internal FPGA. This FPGAenhanced embedded Vision PC also provides two video output ports and an PCIe expansion slot.
Adimec (www.adimec.com), a world leader in application-specific industrial camera solutions, revealed the specifications of the S-65A35 camera created for the future of PCB inspection.
A product demonstrator of this camera with Gpixel sensor was shown at the GPixel booth. This, ultra high-resolution, high frame rate, camera has what the company calls superior sensor characteristics, delivering greater efficiencies and exactness in machine vision applications. The 65-megapixel resolution, CMOS, global shutter camera captures rectangular images of 9344×7000 pixels at 35 frames per second using the CoaXPress interface. This camera is the next step in throughput and accuracy for real time metrology and inspection applications.
With the feature size of both semiconductor lithography and PCB component defects getting smaller the need for increased resolution in inspection cameras becomes crucial. The S-65A35 supports the detail required for smaller technology nodes as well as increased resolution for next generation PCB inspection. With 3.2 micron pixels the S-65A35 puts more pixels per unit area and thus per unit field of view yielding dramatic improvements in resolution, accuracy and precision. In addition, the noise level is significantly reduced, and the quantum efficiency is improved over the existing solutions providing for increased sensitivity and a dynamic range of 70 dB.
Sample availability for the S-65A35 camera is planned for the mid-2019.
The S-65A35 presents high full well capacity of 12.5 Ke-, along with excellent read noise (2 e-) and low dark current. The image sensor and camera together offer high image quality with optimizations available via camera functionality such as defect pixel corrections and uniformity corrections.
The camera will be upheld to the exacting standards of Adimec cameras providing repeatable image performance camera-to-camera for tool matching and consistent product reliability for best cost of ownership. Camera-to-camera consistency and reliability are supremely important. Along with thorough quality control, a critical aspect of repeatability is precise image sensor mounting and alignment to the optics reducing the mechanical and optical alignment time during customer’s manufacturing of their systems. The S-65A35 is also designed to support high mean time between failures (MTBF) for 24/7 operations through advanced thermal mitigation design concepts. The cost of development and system impact is minimized when moving from the Adimec S-25 to the S-65A35, as it has been designed for easy change over to support tool requirements to reduce testing costs and sustained productivity.
Ximea presented their award winning PCIe cameras, USB Vision products and the peripherals that our geared towards embedded, scientific, high speed and high resolution imaging.
For example, XIMEA is expanding the camera lines on offer with multiple new sensors from Gpixel. The CB262 camera model has a high speed PCIe interface to handle the output from Gpixel’s new 26 Mpixel sensor running up to 150 Fps. The Gen3, 8 lane PCIe interface is capable of 64 Gbps speeds allowing the full capability of this sensor to shine through.
The CB654 camera model has the world’s highest resolution Global shutter sensor inside with a small 3.2 μm pixel size and embedded HDR modes. The 65 Mpixel camera delivers 85 Fps for applications where resolution and speed are needed simultaneously. An active EF mount on the camera will allow programmable focus and aperture control for complete authority over the entire imaging chain.
For applications which need very large sensors, the MX377 camera model which was on show will meet all expectations. The 37 Mpixel camera has a 61 x 61mm sensor format for scientific applications like astronomy, x-ray imaging, and electron microscopy.
Both front and back side illumination models are available for the most demanding applications.
Active Silicon had an assortment of its products on display, including a selection of live demos:
- Vision Processing Unit, an embedded computer designed for use in industrial or medical OEM applications, demonstrating live acquisition from four USB3 Vision cameras;
- the latest Camera Link and CoaXPress frame grabbers including our new 4-link CXP-12 CoaXPress frame grabber and the cost-optimized single-link CoaXPress board;
- its frame grabbers in combination with a GPU in a live demo, processing data in real-time;
- its new Harrier series – a camera interface board family which fits neatly onto Tamron, Sony and other auto focus zoom cameras to enable real-time, HD digital video transmission over long cable lengths of up to 700m.
Frans Vermeulen, its Head of Sales and Marketing, delivered a presentation entitled High-speed image acquisition with real-time GPU processing. This was aimed at explaining the pros and cons of processing on a GPU to help you make informed decisions about developing your systems.
Euresys conducted a wide range of demonstrations of its latest frame grabbers, image analysis libraries, FPGA IP Cores, video servers and video converters.
This included the ‘future of machine vision interfaces’ with its brand new Coaxlink Quad CXP-12 which transfers data at no less than 50 Gb/s. It also had a live demonstration of its Coaxlink Octo connected to five high-resolution 4- and 1-connection CoaXPress cameras.
It also showed how Coaxlink is used in embedded applications with demonstrations showing ARM compatibility and its ruggedized Coaxlink Duo PCIe/104 board.
For those interested in innovative image analysis technologies, they could check out the performance of its latest Open eVision inspection libraries: Easy3D and EasyDeepLearning.
There was also the chance to learn how to use the MVDK Machine Vision Development Kit to integrate the USB3 Vision, GigE Vision and CoaXPress FPGA IP Cores from Sensor to Image into your imaging device.
Teledyne DALSA and Teledyne e2v
At VISION, Teledyne DALSA and Teledyne e2v, both Teledyne Technologies companies, combined as Teledyne Imaging to showcase their newest and most advanced vision solutions in a combined display including presentations and live demonstrations.
Key products at the show were:
- CMOS image sensors – new image sensors from Teledyne e2v’s Snappy and Emerald series, including 2, 5, 8.9, 12, and 16M detectors that enable fast and accurate digital processing, and combine full HD resolution with a 2.8 μm low noise global shutter for high-speed scanning;
- Calibir – uncooled long wave infrared imaging and great flexibility in a compact package. Adaptive contrast enhancement, shutterless operation, and VGA resolution with multiple output and lens options;
- Genie Nano 5GigE – new, the industry’s first 5 Gigabit, GigE Vision area cameras with frame rates equivalent to 985MB/sec with TurboDrive for fast, high resolution imaging over long cable distances;
- Genie Nano CXP – a preview of the 25 Mpixel model with the CoaXPress interface offering high resolution throughput equivalent to 2 GB/sec;
- Linea HS – a high-speed, high-sensitivity 16K CMOS TDI camera, with 300kHz line rate and next generation fiber optic interface for high bandwidth, long cable length, low system costs, and high reliability data transmission;
- Linea ML – new, cost-efficient and powerful, Linea ML 8k line scan cameras. Linea ML offers versatile functionalities with mono/HDR, color, multispectral, and multi-field image capture;
- Piranha4 Colour – a 8k tri-linear color line scan camera, built to deliver cost-efficient, high-resolution imaging for color inspection;
- Xtium2 – new, high-performance and fiber-optic enabled frame grabbers support Camera Link HS, CoaXpress (ver 2.0) and Camera Link interface standards. With the Linea ML and Linea HS, the Xtium2 enables the most demanding vision applications;
- Z-Trak – Preview the high-resolution, factory-calibrated laser profiling camera designed for industrial applications requiring object height measurement.
FRAMOS launched their Embedded Vision Ecosystem of sensor modules and adapters.
Designed for a modularized approach to Embedded Vision, the series of FRAMOS products includes sensor modules, module board adapters and processing board adapters for fast and easy development from concept to final design. The FRAMOS EV ecosystem of products provide vision engineers and developers with ready-to-use compositions to speed time-to-market and optimize resources.
Embedded Vision technology often has specific requirements that only a custom solution can solve. The FRAMOS EV ecosystem works as a mix and match tool kit paving an accelerated way building vision products.
André Brela, Product Manager at FRAMOS, explained the concept: “Vision engineers and developers can quickly create a proof of concept, and then evolve it by porting the IP, schematics and board design files into their end product. Built on a very flexible platform, this product line provides capabilities and functionalities to easily evaluate the sensor and its performance on various processor boards, as well as develop the connection and logic components to put all vision-based building blocks together. Developers can work within the ecosystem from the very first idea all the way to mass production, saving resources and shortening development cycles.”
The FRAMOS FSM series of sensor modules is made of pre-existing Sony and ON Semi sensors placed on a PCB with a standardized connector with proper circuit conditioning to get the sensor up and running very fast. The FRAMOS modules also can be ordered with a lens mount and a matching lens. The modules footprint is 26.5 x 26.5 mm, connected via a Hirose 60pin on the bottom side and equipped with 4 holes for mounting the FSM. There are two M12 lens options available as well as a broad range of sensors from VGA resolution up to 12 Megapixel. With the mating FRAMOS ecosystem components, the sensor modules are ready to be integrated and can capture images immediately.
Sensor Module Adapter Boards
No matter what data interface is provided by the sensor, the FRAMOS FSA series of sensor adapter boards can translate it into a common signaling interface through a standardized FRAMOS connector. For greater development flexibility, the FSA adapter boards are designed to work with both FRAMOS and 3rd party sensor modules.
Processor Adapter Boards
The FRAMOS FPA series of processor adapter oards provide the link between the imaging to the processing side. The FPA adapter boards open up the sensor module portfolio to work with a large number of processor development boards, like NVIDIA Jetson TX2, Qualcomm and DragonBoard.
With 37 years of vision experience, FRAMOS supports industrial customers, OEMs, system integrators and researchers overcome the challenges in enabling machines to see. Whether it is about selecting individual components or creating complex vision systems, FRAMOS serves clients with a broad range of components and modular design platforms. Vision solutions can be developed very quickly to help create innovative embedded products and get them to market faster.
LUCID Vision Labs, a designer and manufacturer of unique and innovative industrial vision cameras, demonstrated Helios, a new, compact 3D Time of Flight camera with superior depth precision.
LUCID’s Helios ToF 3D camera has four 850nm VCSEL laser diodes and integrates Sony’s new DepthSense IMX556PLR back-illuminated ToF image sensor with high NIR sensitivity, 10um pixel size and high modulation contrast ratio. The camera can produce depth data at 60 frames per second with 640×480 resolution over a PoE Gigabit Ethernet interface. It is compliant with the GigE Vision and GenICam 3D standard for ease of integration using LUCID’s Arena SDK or third-party machine vision software.
“Our new Helios ToF 3D camera shows the potential for Time of Flight technology in a variety of industrial applications, including robotic navigation, 3D inspection and logistic automation such as bin-picking and package dimensioning,” says Rod Barman, Founder and President at LUCID Vision Labs. “Improved depth accuracy and precision at an attractive price point will enable many 3D applications to run more efficiently while reducing overall system cost.”
The first models of the Helios camera will be available by middle of 2019.
Xenics launched its Manx series at VISION. Based on the XLIN-FC, an in-house developed Iinear detector, the Manx is a high-performance short-wave infrared (SWIR) camera providing high speed and quality line scan imaging. At unprecedented line rates of about 400 kHz, the Manx outclasses the state-of-the-art by a factor of 2.5, becoming the fastest linear SWIR camera available in the world. Achieving new grounds in SWIR imaging speeds potentially paves the way for ever faster inspection processes, increasing daily production capabilities for such applications.
Silicon wafer inspections, TFT screen inspections, food and agricultural produce sorting are some industrial applications that would benefit from this new line of ultra fast linear SWIR cameras. The Manx is offered in 3 different resolutions of 512, 1024 or 2048 pixels. A 2048 pixel version will be available during VISION, demonstrating its high speed, high resolution imaging capability.
The complete portfolio of infrared cameras were presented on the booth. There was a strong emphasis on machine vision applications this year and there were live product demonstrations on how line-scan SWIR cameras are deployed in industries ranging from F&B (food and beverage) to semiconductor.
Research projects can look forward to discovering the SWIR cameras offered by Xenics. Many of these cameras hold unique attributes, such as the fastest area-scan SWIR camera, the Cheetah series, or the Xeva camera series with its extended response up to 2.5 μm. For system integration and product development projects, a wide range of solutions are offered, from ready to use full cameras down to camera modules and detectors.
There were also sneak previews of a next generation camera. The Wildcat 640 is an upcoming SWIR area-scan camera, bringing a combination of high quality and high-speed imaging for demanding industrial applications. The new camera achieves excellent image quality thanks to low noise and dark current. Able to image at speeds up to more than 200 Hz, the Wildcat is well suited for applications such as quality control, laser spot detection, and low light imaging where image quality is crucial. The Wildcat is set for release in 2019.
First Light Imaging
There was a new product for the show, a C-RED 3: High Performance TECless VGA InGaAs camera which is specially designed for industrial applications:
- M600 FPS full frame
- < 50 electrons RON
- VGA 15 μm pixels detector for high resolution
- Electronic shutter <1 μs
- Perfect for short exposure imaging.
- Camera Link or USB3
- Compact and low cost
With the C-RED 3, First Light Imaging has expanded its C-RED cameras family, offering a solution for each of their SWIR imaging needs.
Also showcased during the show was the C-RED 2, its high speed, low-noise InGaAs 640 x 512 SWIR camera offering “breakthrough” performances: 600 FPS full frame and below 30 electrons RON, for all scientific, industrial and biomedical demanding applications.
Allied Vision introduced its first Alvium cameras at VISION.
The company’s show slogan was “Rethink embedded vision,” reflecting a clear focus on industrial embedded vision.
Unique System-on-chip for industrial embedded vision
The Allied Vision Alvium Camera Series is a camera platform that addresses the limitations associated with current camera modules for embedded applications and offers system designers a previously unknown level of freedom. Powered by ALVIUM® Technology, a unique System-onchip designed by Allied Vision, the camera series delivers industrial performance to embedded vision. With a large selection of current sensors, intelligent power management and cost-optimized design, the new camera series combines the advantages of classic machine vision cameras with the advantages of embedded sensor modules.
The Alvium 1500 Series – Embedded vision made easy
The Alvium 1500 Series is the ideal camera to provide easy hardware and software integration for embedded applications. All models come with a MIPI CSI-2 interface which is the most popular interface for embedded vision application development. The Alvium 1500 provides a basic feature set. The software integration can be realized via Video4Linux2 or GStreamer for a simple integration of the camera to the software environment of the embedded system. The configuration of image pre-processing functionalities can be performed directly on the Image Signal Processor in the camera. Depending on the demand of the embedded designers, a Direct Register Access is available to get the full control of the camera.
The Alvium 1800 Series – Best of both worlds
Following both MIPI CSI-2 and USB3 Vision standards, the 1800 Series can be used for either industrial embedded vision or machine vision applications. With an advanced feature set for image correction and optimization, the camera series combines the advantages of classic machine vision cameras with the advantages of embedded sensor modules. It opens up new ways for users to switch from PC-based image processing applications to embedded systems. Designers can choose out of more image pre-processing and advanced image processing functionalities for their embedded vision application, which can be operated directly on the camera. In addition to Video4Linux2 (V4L2) and Direct Register Access the camera control is feasible via the Machine Vision Standard GenICam, providing access to the world of Industrial Machine Vision for embedded vision designers.
New Mako G model with Polarization Sensor Technology
Allied Vision gave its first insights into its new camera equipped with polarization filter technology from Sony.
The new Mako G-508 POL camera is equipped with a PolarsensTM 5.0 Megapixel IMX250MZR/MYR (monochrome or colour) CMOS sensor from Sony. With the new polarizer camera Allied Vision offers a cost-effective tool to unveil features not visible with conventional imaging solutions.
New Ultra-high Resolution Cameras from Allied Vision
Also being showcased were two new cameras for ultra-high resolution imaging applications. The Prosilica GT Large Format camera platform is extended by two ultra-high resolution CCD sensors from ON Semiconductor.
FocalSpec, a supplier of 3D Line Confocal Imaging (LCI) sensors and systems, highlighted its unique technology combining 3D topography, 3D tomography and 2D grayscale intensity imaging at the show.
3D topography imaging is typically utilized in component assembly inspection and online quality assurance in consumer electronics manufacturing. In addition, the 3D topography is suitable for surface inspection and roughness measurement on various types of material which include mirror-like, shiny, glossy and transparent samples with surface angles of up to 20 degrees. In case of higher angles, such as 3D curved glass, the sensor or the sample should be tilted or illustrated with another FocalSpec sensor.
The combination of 3D tomography and 2D intensity imaging can be used for identifying defects such as delamination, scratches or dust on the surface or inside the transparent material. Contrary to various other types of imaging systems, FocalSpec technology helps in not only seeing the place of the defect and identifying in which layer of the sample the defect is. Also, the dimensions including depth of the defect can be detected.
FocalSpec offering for consumer electronics manufacturing
The imaging speed and accuracy vary depending on the LCI sensor most appropriate for the purpose. The most accurate sensor reaches up to 10nm imaging accuracy and the fastest scanning applications perform with the speed of 5000Hz. The amount of measurement points in one measurement line per second is 2048, independent of the sensor type selected.
The FocalSpec range of LCI sensors is available for all third-party developers and systems integrators. FocalSpec 3D Line Confocal Scanner UULA includes ready-made calculation and analysis for the most typical 3D dimensions measurements. The SDK (Software Development Kit) provided with FocalSpec LCI sensors can be utilized by any experienced software developer or integrator for further online application development. Supported software languages include C+, C++, C#, Labview, HALCON, OpenCV and Python. The supported operating systems include various distributions of Linux, and Microsoft Windows versions 7 and newer.
Midwest Optical Systems (MidOpt), a worldwide leader in manufacturing machine vision filters specially designed to enhance contrast and image quality, used VISION to showcase their new Short- Wave Infrared (SWIR) Filters and Wire-Grid Polarizers.
SWIR Filters enhance the image quality of InGaAs camera technology for a variety of machine vision applications and are useful for applications imaging from 900-2300nm. Wire-Grid Polarizers are effective in both the visible and infrared, ranging from 400-2000nm.
MidOpt also showcased the key features of a quality machine vision filter, including wavelength control, passband performance, high transmission, repeatability, and mounting solutions.
Attendees also got the chance to win a free Filter Kit. The MidOpt FK200 Filter Kit featured the ten most popular Bandpass Filters used in machine vision. By using optical filters with white light, the user can determine the wavelength that maximizes contrast and improves the resolution of the feature they want to isolate.
CamPerform-Cyclone, the new high-speed camera series, achieves an overall data transfer rate of 50 Gbit/s thanks to the new CXP-12 interface and four CXP-12 channels. That means Optronis not only offers one of the first cameras to feature a CXP-12 interface, but also an unprecedented data rate.
With the latest 2.0 version of the CXP-12 CoaxPress interface, about to be launched officially, Optronis presented a camera series that utilises one of the world’s most modern data transfer technologies. The Kehl manufacturer has also configured the correspondingly new CamPerform-Cyclone camera series with four channels, enabling a data transfer rate of 50 GBit/s. Designed for 2D/3D AOI and 2D/3D measuring tasks, these camera models offer an appropriate response to the hunger for data of machine vision applications.
“The new CamPerform-Cyclone series meets the consistent demand of our customers for greater bandwidth”, said Dr Patrick Summ, managing director of Optronis GmbH. “In machine vision, where it’s all about details at high speeds, and in high-end automation, where quality is demanded in addition to speed, the CamPerform-Cyclone is an outstanding solution.”
In the CamPerform-Cyclone, Optronis is building on its existing innovations: in 2013 the company presented the first high-speed camera to feature a CoaXPress interface. Now comes the Cyclone, which offers the CXP-12 on 4 channels.
“We can’t afford to relax. The high-speed business, including our niche, demands ever greater bandwidth. Even at VISION, it was already evident that demand for yet more data volume is rising. This makes it abundantly clear that the CamPerform-Cyclone is an important milestone, but one that also points the way in which we want to continue developing”, Summ added.
Optronis announced that its CamRecord Sprinter series is getting a new addition to the family: the Sprinter-FHD. With full HD resolution of 1,920 x 1,080 pixels, the Sprinter-FHD achieves over 1,800 images per second. The camera is controlled via the TimeViewer software, which now also offers functions for object tracking with the current release in order to be able to analyse image sequences automatically.
The Sprinter-FHD in the CamRecord Sprinter series from Optronis has a light-sensitive CMOS sensor and offers full HD resolution at a frame rate of over 1,800 images per second. Various trigger and synchronisation options make the Sprinter- FHD a suitable tool for recording fast processes simply and making them accessible for accurate analysis. The image data is buffered in a fast 16 GB memory, before being saved either to an SSD hard disk internal to the camera, or on a PC.
An object tracking function is integrated into the latest version of the TimeViewer software supplied with every Sprinter-FHD. This allows objects, or points tagged using a marker, to be tracked automatically from image to image. Tracking is at sub-pixel accuracy, enabling precise recording of the path of an object. In addition, changes in angle or velocity vectors can also be recorded. All data are displayed using the software or exported for individualised analysis. Tracking not only simplifies motion analysis in research, but also fault analysis and quality assurance.
“Our CamRecord Sprinter series is well known for good resolution at high frame rates. With the Sprinter-FHD, we will further raise these values. It closes the circle to our new release of the TimeViewer software. With full HD resolution and tracking now integrated, we can run accurate analyses, which promises attractive results in many areas of application,” says Dr. Patrick Summ, managing director of Optronis.
The Sprinter-FHD is available from Q1/2019.
High-speed image processing with 8 GB internal image memory
The Baumer QX series 10 GigE cameras feature 8 GB internal image memory and extended functionalities. Process monitoring applications benefit from inspection times increased by a factor of four as well as flexible usage of the memory. The new function package is available by a firmware update starting the 1st quarter of 2019.
With the expansion of the internal image memory from 2 to 8 GB, more images can be stored, which significantly increases the recording time: 2 s at a frame rate of 335 fps with 12 MP resolution. Flexible memory management, for example 2 MP resolution in a ROI (Region of Interest) nearly provides more than 4s recording time at a rate of 1000 fps. Two further image buffer modes added to Burst Mode enable even more memory flexibility. According to the application requirements, the image memory can be used for a single writing cycle or continuous overwriting like a ring buffer. Further to image buffering, the image preview function simplifies process monitoring – without an HDMI camera interface.
The QX series cameras utilize a 10 GigE Vision compliant interface at 1.1 GB/s bandwidth for fast transmission of buffered images and reduced evaluation time. 10 times faster than GigE Vision and 35% faster than Camera Link Full makes them ideal where short sequences require very high frame rates, e.g. process analysis in industrial applications, live cell diagnostics in the medical industry, scientific research or motion analysis in sports.
The right protection: specially designed housing accessories for the food, beverage and pharmaceutical industries
With the specially developed IP 65/67 and IP 69K housing accessories, Baumer CX.I cameras can now be used to cover applications in the food, beverage and pharmaceutical industries, which set high standards regarding cleaning media or surface roughness in the splash or product contact zone. In thermal terms, the accessories are ideal matches for the cameras and turn them into IP 65/67 or IP 69K cameras in no time at all. This flexibility means that the right version is available for every field of application, with an optimum price-performance ratio. Series production begins in the 1st quarter of 2019.
The IP 65/67 housing was developed in accordance with EHEDG guidelines. Thanks to its round shape, liquids flow off quickly without forming dirt pockets. The hard anodized surface also resists intensive cleaning processes. The washdown design of the EHEDG-compliant, IP 69K, stainless steel housing, eliminates the possibility of product residue adhesion, facilitates residue-free cleaning and prevents the formation of bacteria. With a surface roughness of less than 0.8 μm and foodgrade, hard-wearing seals, it is also suitable for splash and food zones in which oils and greases or chemically aggressive cleaning media are used. The housing kits are supplemented by an extensive range of tubes to meet the various application requirements. Tubes are available with different cover glass materials such as acrylic glass or chemically strengthened alumino-silicate glass with high scratch, impact and fracture resistance as laminated safety glass for a long service life.
The CX.I cameras with resolutions of up to 12 megapixels and up to 148 fps can withstand shocks of up to 100 g, vibrations of up to 10g, and have 4 power outputs with pulse width modulation and a power output of up to 120 W (max. 48 V / 2.5A) to control lighting directly without an external controller. Models with an extended operating temperature range from -40 °C to 70 °C also allow for use in extremely demanding ambient conditions.
LX series cameras with 10 GigE enable the control of autofocus liquid lenses for instant and flexible focusing at varying working distances or object heights. Combined with advanced features like shading correction and IEEE 1588 capability, they easily apply themselves to new applications in measurement technology, logistics and the pharmaceutical industry and flexibly adapt to individual application requirements. A firmware update to easily upgrade previous LXT models will be available in the first quarter of 2019.
Electrically focusable Corning Varioptic liquid lenses excel in applications with quickly changing object distances that require fast autofocusing e.g. in 2D code inspections on packages of different heights. The autofocus lenses are powered and controlled via the camera connector. The image processing system focuses within fractions of a second, even between largely varying working distances and without mechanical movement or compromise to image quality or speed. This simplifies and speeds up many applications within a production line.
The 10 GigE LX cameras have many new powerful functions: auto features to adapt exposure times and lighting to the prevailing light conditions, sensor multi ROI to increase frame rates and reduce the amount of data transmitted, as well as shading correction to correct lens and lighting artifacts. In addition, the cameras support IEEE 1588 for precise time synchronization in multi-camera systems and RS232 for communication with external serial devices.
The LX series includes more than 40 models with Dual GigE, Camera Link or 10 GigE interface and cutting-edge global shutter CMOS sensors by ams, ON Semiconductor and Sony. With resolutions of up to 48 megapixel and frame rates of up to 337 fps, they are the ideal choice for inspection tasks with demanding requirements on image detail and throughput.
Getting top billing on the Gardasoft booth was a new dual-channel lighting controller which has been designed specifically for the precise control of OLED panels in machine vision applications. Featuring sophisticated drive techniques to meet the complex physical properties of OLED panels, the new controller allows the generation of extremely fast, highly accurate and stable light pulses at the limit of the OLED operation. In addition to the announcement of a new lighting controller, Gardasoft also presented a number of new demonstrations to illustrate the power of machine vision sequencing and control technologies. These will cover computational imaging, ultrafast focusing, OEM controllers and multi-light line scan stations. The new lighting controller will open up new illumination possibilities that will further contribute towards novel machine vision solutions.
VISION provided the first chance for people to see the application of Gardasoft’s RTCC series of lighting controllers for high-speed computational imaging. The controller will be used to drive a segmented dome light to allow a rapid sequence of images with different illumination angles to be acquired by a smart camera. The camera will use computational imaging software to extract surface detail that cannot be seen using conventional imaging methods.
Three controllers were in action to demonstrate high-speed focusing on objects of different heights using a liquid lens from Optotune AG. A CC320 timing controller, positional encoder from a moving platform, camera and RC120 lighting controller will be linked together to give the required camera exposure and illumination needed for imaging. A TR-CL-180 liquid lens controller will be used for high speed adjustment of the lens for focusing. In addition there will be the chance to explore control options for OEMs through the board level TR-100 and TR-CL-190 lighting and liquid lens controllers.
VISION also provided the opportunity to see see how multiple lights can be used with line scan cameras for the capture of image sequences under different lighting conditions. Information from different LED sources is captured on sequential lines and individual images for each source extracted using software.
At VISION, IDS Imaging Development Systems demonstrated the full potential of IDS NXT. The vision app-based industrial camera platform is expanded by the IDS NXT rio and rome product families and is offered with additional sensors and communication interfaces. For the first time, the company integrates neural networks with selflearning algorithms within some of these models.
In the field of 3D vision, the company presented Ensenso XR, the first stereo camera of the series that can calculate 3D point clouds itself. Novel focusable USB 3.1 Gen 1 board level industrial cameras with liquid lens control were on show at exhibition stand 1F72, together with concept studies of extra fast 10 Gigabit Ethernet and uEye cameras with polarization sensors as well as numerous live demonstrations.
The new IDS NXT rio & rome families are designed as intelligent cameras in the housing of standard industrial cameras. This approach combines the capabilities of both worlds: customers can use them as fully- fledged standard industrial cameras, or take advantage of extended, individually designed applications based on vision apps and on-camera neural networks.
For greater flexibility regarding 3D vision applications, the company presented a prototype of the new Ensenso XR series with on-board processing. Unlike the cameras of the N- and X-series, which require a host computer to calculate 3D point clouds, this model can calculate them itself and transmit the data via Ethernet or Wifi. The advantages include faster generation of these 3D point clouds, less strain on network bandwidth due to transmitting results and a reduced computing load on the host PC. This allows faster clock rates for bin picking, for example.
In addition, IDS presents the concept study of a 10 GigE camera family that could make best use of the potential of current sensors in terms of high resolution, fast frame rates and large bit depths. Depending on the available bandwidth and infrastructure, it automatically adapts its transmission speed to 5, 2.5 or 1 GigE. The company also plans to integrate Power-over-Ethernet (PoE), which means that a separate cable for the power supply is no longer required. The models could be particularly interesting for applications that could only be operated at a reduced bandwidth due to the limited bandwidth of “standard” GigE cameras.
Various demonstrations emphasized what the different products in the camera manufacturer’s portfolio are capable of. For example, IDS showed the new USB 3.1 Gen 1 board level cameras with liquid lens control. They facilitate image acquisition at variable object distances as their focus can be readjusted quickly and comfortably via user interface or API, even if the lens cannot be reached manually. The cameras are available with a 6.4 MP rolling shutter sensor from Sony or the highly photosensitive 18.1 MP rolling shutter sensor from ON Semiconductor, among other variants. They are equipped with S-Mount or CS-/C-Mount, twist-proof USB Type-C connection and practical USB power delivery.
The IDS team also provided GigE Vision camera prototypes with the 5 MP IMX250 Pregius CMOS sensor with integrated on-pixel polarizer from SONY. The polarization sensor provides better object detection in low contrast or reflective light and helps to visualize scratches on surfaces or stress distribution within transparent objects. The conversion of the on-pixel polarizers is already realized in the camera. As a result, the camera delivers the result image directly, CPU-intensive evaluations of the polarization on the host PC are not necessary.
LMI Technologies, a global leader in 3D scanning and inspection, unveiled three new 3D inspection products at VISION.
Gocator 2500 Series Smart 3D Laser Line Profilers
Gocator 2510 and 2520 are the most advanced Gocator line profilers to date. Designed for highperformance small parts inspection, these highspeed sensors generate precision 3D scans down to 8 μm X resolution. With a custom high-speed imager, advanced optical design, dedicated FPGA processing, and built-in measurement tools for 3D alignment, part segmentation, and 3D feature extraction, Gocator 2500 sensors handle the complete processing pipeline––from raw image data to 3D result––in an all-in-one device that operates at inspection rates up to 10kHz.
Gocator 3504 Smart 3D Snapshot Sensor
At 6 μm the Gocator 3504 is the highest XY resolution 3D snapshot sensor on the market today. Combine that with +/-0.5 μm measurement repeatability, and this groundbreaking sensor provides robust inline factory automation for sub-micron level inspection.
The 3504 has an industrial stereo camera design and point-and-shoot technology inside a ruggedized industrial package that allows for highly accurate inspection without a motion stage. The sensor is pre-calibrated and ready to measure out-of-the-box using just a web browser.
The Gocator 3504 sensor can be used in a wide range of electronics and medical applications such as connector and pin coplanarity, wire detection, surface flatness, and stent inspection.
GoMax Smart Vision Accelerator
GoMax is a cost-effective hardware solution that allows users to accelerate any Gocator® 3D smart sensor in order to meet inline production speed, without the need for an industrial PC. This plug-and-play approach connects directly to any Gocator® model and speeds up the sensor by taking over a portion of the data processing (including point cloud generation, 3D measurement, and PLC/robot communication).
“The launch of these products continues to advance our FactorySmart approach to automated inline quality control, automation, and optimization. By offering new benchmark products in three key technology areas, LMI is serving the full spectrum of machine vision customers and working to continually meet their specific application demands,” said Terry Arden, CEO, LMI Technologies.
Press Conference bonanza
VISION is an ideal time for press conferences and, apart from the organisers main event (which I attend, because they give you breakfast), I managed to take in a couple of conferences.
EMVA took us through a new EMVA Standard Initiative: Embedded Vision Interface Standard.
It explained that cameras and PCs as the main components of machine vision systems have tremendously miniaturized over the last years. The combination of a processing board with a powerful small camera, make it possible to design a very compact vision system which can be integrated into a larger system. Such systems are called embedded vision systems and are of high interest to the machine vision industry. However, adaptation is needed in order to enable industrial solutions to use embedded systems. Hosted by the EMVA, in August 2018 a new standard initiative formed to address these needs.
Embedded systems for industrial solutions
With impressive support from the machine vision industry, the standardization group develops a standard to achieve an easy integration and exchange of different embedded cameras within embedded vision applications. Already more than 40 companies are showing interest in this initiative and the working group of the standard includes delegates from Adimec, Allied Vision, Alysium, Avaldata, Basler, Baumer, Euresys, Flir, Framos, Matrix Vision, and Sony Semiconductor Solutions Corporation.
Using reusable modules in kernel-mode, developers of embedded vision systems could reduce integration efforts, development time and, as a result, costs. The needed components such as light, lens, camera and computer could be integrated and connected to embedded boards more easily.
Another goal is to give embedded system designer access to the machine vision standard GenICam (Generic Interface for Cameras), a generic programming interface for all kinds of devices. With this standard they would also benefit from the advantages of an identical application programming interface (API) and the easy-to-use GenICam modules, regardless of the interface technology. In addition, standard group intends to define GenTL as the main kernel-to-user interface.
Another important aspect the Embedded Vision Standard is intended to solve is the enhancement of the sensor interface standards SLVS-EC / MIPI CSI-2 D-Phy with recommendations for hardware components, like cables and connectors.
The next steps of the standard initiative are a white paper with concepts and goals to be adopted early next year. A first release candidate is targeted in 2020.
New EMVA Standard Initiative: Open Lens Communication Standard
Amazingly, there is one component in an image system that has not yet changed since the early days of the machine vision industry: the lens mount. C-mount is the most common type of screw lens mount and much older than the machine vision industry. In the constantly changing world of machine vision, the lens mount seems to be something rock solid. But this also hindered progress. Already in 1987 Canon introduced the Electro-Optical System (EOS) for digital information and power transmission between SLR cameras and lenses, and used it to produce lenses with built-in autofocus motors. Such a standardized system is still not available for the machine vision industry.
Take all the possibilities that have already been implemented in commercial system cameras plus modern methods of computational imaging, and combine these with the processing capacities of modern embedded vision systems. Only then one gets an idea of how powerful and different the next generation of image acquisition systems could be and what the machine vision industry still misses with an open standardized lens-to-camera communication.
This includes auto-focus systems (also with liquid lenses), setting the focal length for zoom lenses, control and reading the aperture setting, and inquiry of lens properties by the cameras, such as aperture dependent lens shading, geometrical distortion, and lateral chromatic aberration. With this information the camera can automatically correct these distortions. Automatically capturing a focus series, and computing depth maps and images with extended depth of field and correction or modification of the modulation transfer function (MTF) of the lens or camera system are also possible.
This is why the EMVA has decided to start a new standardization group on an open lens camera communication standard, which may have different mechanical connections but a common protocol closely linked to the GenICam Standard. A first meeting of the standardization group took place on 9 July at Heidelberg University with an initial setup of the group. Marcel Naggatz from Baumer Optronic and Erik Widding from Birger Engineering were elected as Chair and Vice-Chair of the standard initiative. Already, major camera and optics manufactures have joined the working group. The second meeting took place on 3-4 December in Radeberg, Germany.
With a billing that they will change the industry forever, this press conference had a lot to live up to.
And in the end, after a sluggish start, it did. Whether Inspeckto have got it right, remains to be seen, but it’s proposition is compelling.
At the show it launched the S70, the world’s first Autonomous Machine Vision system. The S70, said Inspekto, offers powerful quality assurance capabilities in a small, versatile and practical package. Suitable for any handling method, product type and material, the S70 is up to the job, no matter what the visual QA task.
The S70 system can be installed in 30 to 60 minutes, 1,000 times quicker than a traditional machine vision solution and at ten per cent of the cost. This out-of-the-box system offers a simple, intuitive user interface designed to be installed directly by the shop-floor employee. This means that no systems integrator is required at any step of the short set up process, and at any time later. The S70’s high affordability and ease of deployment enable manufacturing plants to install it at any point on a production line, and even move it from one line to another, at any time in the future, within minutes.
A German company with Israeli DNA, Inspekto is supported by leading industrial businesses from across the DACH region. During beta stage, the company installed its system in the plants of leading industrial manufacturers, in countries including Germany, Italy, France and Austria. The S70 will deliver market changing benefits to manufacturers with a yearly total available market exceeding $30bn.
“The S70 is a world first – defining the Autonomous Machine Vision category and introducing the inaugural Plug and InspectTM technology for the modern shop floor environment,” explained Harel Boren, CEO and co-Founder of Inspekto. “Because of the S70’s affordability and simplicity, the digital factory is now a reality, allowing collection of data, down to product-images, meta-data and defects from the entire production process and across production tiers. It offers full archiving and traceability – protecting both the manufacturer’s production process, as well as its customers, from unwanted scrap and unwanted defects.”
“The market has been waiting for the arrival of the S70,” added Zohar Kantor, VP Sales of Inspekto. “Manufacturers are trapped into expensive contracts with systems integrators and cannot access machine vision technology themselves. Inspekto’s S70 system finally puts the manufacturer at the center-stage. Our powerful, and growing numbers of world-leading customers attest to the huge impact of the S70 in responding immediately to any QA need arising on the production line. There is no need for any external experts, there’s no need to select cameras, lenses or any other equipment required with a traditional solution, there’s no need to take products off-line for inspection and there’s no need to put any special structures in place. By setting their visual QA systems, directly by themselves, the S70 gives manufacturers independent control on their QA, first time ever.”
“Our team of research experts have pushed the boundaries of computer-vision and artificialintelligence technologies to develop a product that will change the QA industry forever,” added Yonatan Hyatt, CTO and Co-founder of Inspekto. “Working alongside some of the world-leading manufacturers across the DACH region, we have developed an autonomous product introducing an innovative and patented technology that dramatically increases the productivity of production and QA managers. It truly meets the challenges of today’s manufacturing world.”
Inspekto is working with a handful of leading German investors, including such top players as Mahle, Grazia, Steinbeis and ZFHN, and numerous leading DACH, French and Italian top global manufacturers, to revolutionize QA in manufacturing.
Inspekto’s Plug & Inspect technology means that employees in the manufacturing facility can take the Autonomous Machine Vision system out of the box and install it in minutes, on their own, without any help from a systems integrator and without the long process associated with setting up QA solutions today. Plug & Inspect powered systems provide full autonomy to the manufacturing plant: systems immediately self-adapt to any changes in the inspected object environment, such as changes in light conditions and object location or orientation, without any intervention whatsoever.
The system’s powerful machine vision and artificial intelligence capabilities mean that manufacturers can also independently change the object being inspected and the QA system, or even move the system to another production line altogether, on their own, in minutes, with no need for any external experts. “Traditional QA methods rely on systems’ integrators to install and adjust the QA system,” explained Harel Boren, CEO of Inspekto. “This imposes a tedious expert-dependant setup and commissioning process and ties manufacturers to a third-party installer, leading to inescapable excessive costs and downtime. Autonomous Machine Vision changes this. It makes the systems’ integrator entirely obsolete and brings control in-house, giving manufacturers the opportunity to plug and inspect in minutes. We put Inspekto on course to change machine vision QA from the necessary evil it is, to be customercentric, and give QA managers a tool that is 100 times quicker to set up, at 1/10th the cost.
“VISION is the best place for us to hold the ‘world premier’ of the industry’s first ever Autonomous Machine Vision system, powered by our Plug & Inspect technology,” Boren continued. “We will be running live systems on exhibition grounds so that manufacturers can experience first-hand how the new category of Autonomous Machine Vision systems will transform the way they perform QA.”
The robust technology behind Inspekto’s product, leading to unprecedented affordability and immediacy of installation, finally makes Total QA possible for any industrial plant, whereby inspection is conducted at every stage along a production line and is no longer limited to mission-critical points or the end of the line.
Total QA allows quality defects to be identified earlier than with traditional QA methods, which reduces the amount of resources wasted on the manufacturing of objects with quality defects and carries vast direct impact to industrialists’ bottom line.
Harel Boren, CEO of Inspekto, the founder of Autonomous Machine Vision, explains the truth behind some common machine vision misconceptions.
“The introduction of the USB drive led to the eventual decline of the floppy disk, which is now banished to the history books of storage. As the machine vision industry goes on to embrace Autonomous Machine Vision, misconceptions remain.
‘Cost limits what’s possible’
“Installing a machine vision solution is viewed as an expensive process. This is because traditional machine vision solutions require a major effort and continuous investment on top of upfront costs. Installing and maintaining a traditional solution is a complex process that requires expertise. The combined cost of these professional services with cameras, lenses, lighting and more, turn traditional solutions into an expensive business – in the region of €20,000 to €150,000 or more, per inspection point.
“Because these solutions are tailored to a particular point in the production line, they are also inflexible, made to work for only one product at one point on one production line. This means adding additional quality assurance points to the line can cause costs to spiral.
“None of this is true with Autonomous Machine Vision, where a system is standalone, simple and quick to setup – making it affordable. In fact, an Autonomous Machine Vision system should typically be installed at 1/10th of the cost and at up to 1,000 times the speed of planning and installing a traditional solution. Thanks to speed and affordability, it can be installed at any stage of the production line – and moved to a new location in minutes when required.”
‘A systems integrator is essential’
“Because traditional machine vision is a complex process, there is a misconception that the industry will always rely on the expertise of a systems integrator. The machine vision ecosystem has for many years aligned itself to the systems integrator, the only party equipped with the expertise needed to build a solution. Once the solution is installed, a systems integrator is required for every set up and changeover on the production line.
“In the age of Autonomous Machine Vision, systems integrators are no longer required. The manufacturer can install a visual quality assurance system out of the box in minutes. Because Autonomous Machine Vision systems can self-set and self-adjust, the manufacturer is able to do this independently at any point on the production line. Just like the floppy disk, the systems integrator becomes obsolete.
‘Machine vision will never be plug and play’
“Since the 1980s, manufacturers have used machine vision technology for quality assurance. The process has always involved building a tailored solution, piecing together the right filters, lenses, lighting and cameras. Over weeks or even months, the integrator must continue setting proof of concepts (POC), testing plans, programming and more – leaving plug and play visual QA as a far-fetched dream. As written by Fred Grootentraast from ICT Group in a 2016 article titled Machine vision is not plug and play, “The success of machine vision is dependent on the use of the right hardware and software and this requires specialised knowledge of both disciplines.”
“Though this was once true, Autonomous Machine Vision now brings the capability to simply Plug and Inspect. New technology, such as Inspekto’s S70 system, mean that in 30 minutes, any member of staff could install a system and start performing effective visual QA.
“Just as the next step in storage was not a smaller floppy disk, but an entirely new approach, Autonomous Machine Vision is a new era in visual QA.”
Leading multinational industrial companies from Germany and Switzerland have invested more than $10m into Inspekto. The start-up aims to meet the needs of a yearly multi-billion-dollar market across a broad range of vertical manufacturing sectors, including automotive and electronics.
The $10m injected comes from leading industrial players, as well as expert financial investors. Inspekto’s investors include: Grazia Equity, ZFHN, Mahle, Planven, THI Investments and Steinbeis. The company’s valuation post investment is in the region of $60m.
The funding is for Inspekto’s launch activities of Plug & Inspect technology driving its leading product, the INSPEKTO S70. During research and development (R&D), Inspekto collaborated with tens of leading DACH manufacturers to ensure the product is perfectly aligned with market needs and requirements. The company is now heading to its official launch at VISION 2018 in Stuttgart, where it will hold live product demonstrations and make its product available for general sale to manufacturing plants worldwide.
“The INSPEKTO S70 is the world’s first ever Autonomous Machine Vision product, which will reinvent machine vision QA,” explained Harel Boren, CEO of Inspekto. “Leading players chose to invest in Inspekto because of its huge expected impact on industry,” added Boren. “This financial support offers a vote of confidence in the expected impact of Autonomous Machine Vision on the quality assurance (QA) industry – a multi-billion-dollar market. Inspekto’s mission is to make quality assurance managers love their jobs, by giving them complete control of where, when and how visual QA can be conducted.”
“Inspekto’s research team aspires to explore and push the boundaries of computer-vision and artificialintelligence to allow autonomous machine-vision solutions for industrial manufacturers,” added Yonatan Hyatt, CTO of Inspekto. “Our introduction of zerofriction setup of inspection units throughout entire production lines is bringing factories to peak levels of efficiency and ensuring immediate detection of defects right as they appear in the manufacturing process. The current financing will allow us to further deepen our domain-specific research, alleviating major points of pain in the industrial domain.”
“As one of the largest automotive suppliers worldwide, we identified early-on the enormous potential impact of Autonomous Machine Vision on production, performance and competitive edge,” said Johannes Diem, Corporate Planning, Corporate Strategy, Venture Capital and M&A Strategy at MAHLE International GmbH. “Inspekto’s work with our plants has been consistently useful and led us to make our first-ever direct investment in a start-up company.”
“Steinbeis has a long history of working closely with German industrial manufacturers,” explained Uwe Haug, Member of the Board of Directors of Steinbeis GmbH & Co KG für Technologietransfer. “We identified early on the strategic impact that Inspekto brings to German industry, across any vertical and vendor size, which is what drove us to make our investment in the company.”