A preview of the entrants for the Vision Award, sponsored by Imaging and Machine Vision Europe, presented each year at the Vision Show in Stuttgart (9-11 November). The prize recognises outstanding products or processes in the area of applied imaging, with the winner receiving €5,000
The M200CT from ABAQuS is designed for quality control of 2D barcodes. It is an easy and competitive test system providing compliance to different engineering standards.
With the Prosilica GX, Allied Vision Technologies introduced a double-speed GigE Vision camera using Link Aggregation Group technology (LAG).
To achieve a rate of 240MB/s the GX-Series uses Link Aggregation Group (LAG), or IEEE 802.3ad, a networking technology that uses multiple Ethernet ports in parallel to increase the link speed. The camera features two Gigabit Ethernet ports configured as a LAG. When the camera is connected by two cables, the host computer only sees it as one connection at twice the normal speed (240MB/s).
The GX-Series can also work at half the speed (120MB/s) using a single cable. The LAG technology has been used in IT networking for years and is supported by a wide range of standard Ethernet hardware. The Prosilica GX operates with standard CAT5e Ethernet cables and only requires an LAG-compatible interface card on the PC-side.
SAL3D enables complex 3D applications in the production line for 100 per cent parts inspection. Freeform parts moving on a conveyor belt can be accurately scanned, obtaining a reliable metric 3D point cloud for each of them, using the Peak Detector and Metric tools from SAL3D, optionally using the Merger tool for minimising occlusions.
The Match3D tool will obtain the pose of each object related to a known model. It will provide the means for geometrically transforming the parts point cloud to a well known pose. Here, either a global shape measurement can be made through comparison with the model, or different local 2D measurements of freely oriented flat features, such as circles, can be performed by finding their corresponding projections onto known planes, generating a ZMap for each of them. These ZMaps can then be fed into 2D tools in order to obtain the circles' diameters.
Either as a standalone package or in combination with 2D tools, SAL3D enables 3D complex and accurate measurements on parts in arbitrary orientation.
BAPis has developed a high-end, high-speed scanner. It uses a CIS sensor and has the ability to scan documents up to DIN A3 with a scanning throughput that is competitive to a scanner based on CCD sensors.
Due to re-scaling on the CIS interface board and controlling the speed of the transport system, it is possible to scan not only 600dpi, but also 300, 200 and 100dpi resolution of the images. Lower resolution requires higher speed of the transport. 200dpi images of DIN A4 paper size can be scanned with 1.2m/s, which means it is possible to reach up to 240ppm. This was made possible as the sensor is equipped with an integrated LED white light source and fast processing on a BAP image processing board. BAPis developed a scanner that is much faster than other scanners basing on CIS sensors.
Basler's innovation is the combination of several different existing technologies into a new camera generation. Basler will launch it as part of its Ace camera series. The first model is named acA2500-14gc. This camera will be the pioneer for additional models based on the same concept.
The new camera combines several distinct aspects. It unites an extremely compact 29 x 29 x 42mm housing with Gigabit Ethernet interface technology, power supplied via Power over Ethernet, and a 5 Megapixel CMOS sensor that can deliver 14fps thanks to the GigE interface. In addition, it will have innovative algorithms that result in a remarkable improvement in colour performance. The uniqueness of this new camera generation resides in the combination of all of its facets. With these novel new products, Basler will revolutionise current price levels for fast, high resolution cameras.
Drawing from over two decades of experience in video eye tracking, Chronos Vision has developed a novel high-speed, real-time video eye tracker, based on the Sick/IVP Ranger camera.
On-chip parallel processing and the DSP/FPGA architecture reduces latency between image acquisition to data output and enables high sampling rates. An image sampling rate of 1,350Hz with a latency of 600 microseconds is achieved.
Refined algorithms track eye features (pupil, limbus) regardless of image artefacts, while synchronous acquisition from two or more regions-of-interest facilitates binocular eye and head recording.
One of the main applications for the eye tracker is during refractive eye surgery, in which the combination of high spatio-temporal resolution and short latency ensures accurate, real-time measurement of even the fastest eye movements and permits accurate control of the surgical laser.
Furthermore, the high spatio-temporal resolution and, more critically, the short latency are relevant to numerous machine vision applications where accurate, real-time object tracking is required.
BOA is a highly integrated smart camera comprising all the elements of an industrial machine vision system. It incorporates multiple processing engines, which enables algorithm optimisation via DSP, application management via CPU, and sensor management via FPGA. Dalsa claims it is the first smart camera that offers truly embedded application software, which is easily set-up through a standard web browser. This means there is no software to install and therefore no version control problems.
Specifically designed for industry, BOA's 44mm cube form factor is perfect for tight-fit applications. The IP67-rated housing means that the camera can be directly deployed in harsh, wash-down environments. This is particularly useful to meet cleanliness standards in the food and pharmaceutical industries, and eliminates the need for a separate, costly protective enclosure. BOA's small form factor and easy mounting capabilities allow it to integrate easily into existing production lines, machinery or moving equipment.
Effi-Lase is a LED lighting range specifically designed by Effilux for 3D reconstruction by profilometry. This LED lighting device projects very thin and powerful light lines for applications such as 3D reconstruction (semiconductor, food and beverage, solar, mining, and pharmaceutical industries) or positioning (alignment, targeting and positioning applications).
Effi-Lase can be compared to a laser but without all the drawbacks. Indeed, it uses LED technology and can project lines without speckle, with perfectly sharp edges, high power, excellent illumination uniformity and no ocular hazard.
This product has been developed to be usable by everyone with easy installation, little maintenance and a very long lifetime. It is flexible, with an adjustable working distance and adjustable targeting area.
Flir's A615 is an uncooled infrared camera with a 640 x 480 focal plane array. It is fully compliant with both GenICam and GigE Vision protocols. This combination makes this high resolution, high frame rate camera ready for quick and easy network installation, enabling the user to take full advantage of plug-and-play with machine vision software from National Instruments, Cognex, Matrox, MVTec, Stemmer Imaging and many more.
The introduction of the fixed-mounted Flir A615 gives manufacturers and system integrators even more reason to choose infrared over a visible light camera for high-end automated inspection and process control. Its ability to provide accurate temperature measurement with high quality, high speed imaging gives the camera the edge for a wide range of more dynamic vision applications.
Frankfurt Laser Company offers a new line of high power, temperature-stabilised laser diode modules. The HEML series combines all of the requirements for industrial machine vision applications.
The unit has been specially designed for use within harsh environments. It provides users with a laser light source that can operate continuously in difficult conditions. The module is developed for use outside laboratory or clean working environments, allowing laser light to be used in places where standard laser diode modules would not operate well or for prolonged running periods, both requirements of industrial manufacture. By combining the feedback from our customers and the company's engineering skills, Frankfurt Laser Company has created a unique product within the market.
Eye tracking for machine vision is a very demanding task regarding accuracy and speed. A positioning error of the pupil of less than 1mm is required and capturing fast saccades at 1,000fps is necessary. Nowadays eye trackers are proprietary, expensive and cumbersome to use as they require a tedious calibration procedure before image processing.
The new eye tracker solution from Fraunhofer IDMT with a scalable and configurable Hough IP core overcomes all these obstacles. There is no calibration step necessary. Depending on the limited image size the eye tracker can process frame rates of more than 1,000fps and fulfils all requirements regarding the positioning error.
In this paper the Inviso ImageBoost platform is presented, which is a novel image processing platform inspired by the human visual system. The human visual system is very robust and capable of handling a wide variety of lighting conditions. One important reason for this is because it uses contextual information of the image at a low level.
At the Eindhoven University of Technology the orientation sensitive cells of the primary visual cortex are mathematically modelled and this model is implemented by Inviso on an FPGA resulting in the Inviso ImageBoost platform. This platform can be used in a wide variety of applications delivering high speed and accurate detection under hard lighting conditions. Some examples include crack detection in natural stone slabs, line enhancement and quantification of fibre alignment in medical imaging, distance measuring between semi-transparent sheets and ball detection for soccer robots.
The Keyetech Texture-based Recognizer is a software package for rapid pattern matching and 2D/3D object localisation using local point features. It is robust to rotation, perspective skew, scale changes, occlusion, and changing lighting conditions.
The software provides high-speed feature computation: Keyetech's highly optimised CPU implementation of the Harris-SIFT features developed by Dr Pedram Azad exploits two levels of parallelisation: SIMD instructions (Single Instruction Multiple Data) and multicore technology.
With local point features, the most time consuming part becomes feature matching. Therefore, Keyetech offers optional GPU acceleration, which allows frame rates of up to 100Hz. In addition, the Keyetech Texture-based Recogniser is implemented within the framework of the Integrating Vision Toolkit (IVT) - an open source computer vision library, which is maintained by Keyetech in cooperation with the Karlsruhe Institute of Technology (KIT).
New Imaging Technologies (NIT) offers CMOS imaging sensors based upon a unique and patented pixel technology, which provides intrinsic high dynamic range response of more than 120dB, no noticeable fixed pattern noise and operability without image artefacts to more than 90°C.
NIT's Native WDR (Native Wide Dynamic Range) sensors are based on a new generation of CMOS imaging technology named Magic (Matrice Active à Génération d’Image indexée sur Contraste), capable of generating a highly stable image even under very wide intra scene dynamic range (>120dB). This is in a single frame time and without multi-exposure, any setting of knee points or external control.
NIT offers a complete range of standard, customised and full custom CMOS imaging sensors for a variety of applications where performance, cost, quality and reliability are key requirements.
OPT Machine Vision Tech's AOI light application consists of three LED arrays, which have different colours and radiation angles. This structure can map the slant surface shape of the components, with different colours representing different slant surfaces. The system can therefore judge the situation of the slant surface by the colour reflection.
The AOI illumination system can be the basis of various types of defect detection, for example, salient point, pits, scratches, all of which are more difficult to detect under standard illumination. The lighting system is ideal for PCB inspection, such as PCB assembly inspection in surface mounting technology, especially for detecting defects of solder joints, and inspection of the pins of ball grid arrays, a type of IC packaging.
Photometrics' Evolve EMCCD camera is engineered to report data in standardised units, photoelectrons, in real time. With low-level programming access to the CCD pixel array, the Evolve camera can immediately provide any machine vision application with standardised imaging data without any calibration budget, equipment or expertise.
The Evolve was designed for sensitive applications, such as LED and photovoltaic cell inspection or spectral imaging of pharmaceuticals, in which calibration requires significant time, budget, specialised equipment and human resources. Standard cameras require continuous calibration because of the way they amplify photoelectrons, report data, which varies between units, and as sensors age. Photoelectrons, produced directly by light incident to a camera's sensor, are independent of the camera or user-selected settings and therefore enable direct, quantitative comparisons of generated data.
PMD sensors from PMD Technologies enable the real-time measurement of depth maps and grey-value images at lateral resolutions of up to 204 x 204 pixels according to the time-of-flight (ToF) principle.
This technology provides new perspectives for human-machine-interaction: since the depth map measurement is independent of texture, environmental lightning conditions or gloved hands, the segmentation, detection and analysis of human hands can be accomplished in a way superior to the capabilities of 2D image-based approaches. With the CamBoard, PMD Technologies provides a fully USB-powered, webcam-sized ToF camera with an excellent potential to be integrated into consumer products. It enables intuitive and natural ways of hand-gesture based interactions with computers, machines, mobile devices, etc. PMD Technologies has set up a trial implementation of a robust hand-gesture interface providing mouse cursor positioning, clicking and drag-and-drop interaction.
Raytrix has developed camera technology and software algorithms that, from a single snapshot, can computationally generate images with an extended depth of field, variations in camera viewpoint, stereo views of the scene, and an estimation of 3D scene depth.
The technology is based on a micro lens array (MLA) with a specific design developed by Raytrix, which is placed at a precise distance from the image sensor. The price that has to be paid for the additional features is a reduced effective resolution of at most one quarter of the sensor resolution and a varying effective resolution over the depth of field. The highly parallel algorithms developed by Raytrix that perform the refocus calculations can generate approximately a 1 Megapixel image in around 10 milliseconds. The cameras are available directly from Raytrix or from IDS starting at around €1,000.
Moritex and Schott have introduced a telecentric HR 6x zoom lens for 5 Megapixel sensor cameras. The ML-Z07545HR lens makes it possible to explore the entire potential that 5 Megapixel sensors offer. With its optical zoom (0.75x up to 4.5x) that can be operated either manually or electrically, this lens can be used in a variety of different ways. The iris diaphragm and focus can be adjusted with high precision. The compact, lightweight lens is equipped with a C-mount. This product series offers integrated coaxial lighting that allows for even illumination of the object field.
Its long working distance of 70.9mm makes it less susceptible to heat, particles or oil during inspections. Episcopic lighting can be integrated upon request for more homogeneous illumination of objects under examination. Moritex and Schott offer high-performance LED and halogen light sources that come equipped with fibre optic light guides on an optional basis. A wide range of accessories are also available.
Sick is presenting a high-speed 3D camera with simultaneous high-quality colour capability at up to 3,072-pixel resolution. The ColorRanger E combines the functionality of a 3D camera and a colour line scan camera. It can be configured to provide several different measurements at the same time, such as 3D shape, colour, laser scatter, and monochrome, according to the exact need of each specific application. As a result, one ColorRanger can replace the needs of several different cameras, which reduces the system size as well as the hardware and integration costs.
ColorRanger E is based on a unique CMOS sensor with on-chip processing and software controlled windowing. The 3D data is based on laser line triangulation. Examples of applications where both colour and 3D shape matters can be found in various industries such as food, cosmetics, pharmaceutical, electronic assemblies and wood.
Smartvision has developed a tool, BlobMax, for the inspection of planar and curved (also cylindrical) surfaces. It is ideal for inspecting surfaces that show diffuse or specular reflection behaviour.
The tool inspects the entire visible area of planar and cylindrical surfaces, including the border of objects. A brightness gradient along the curvature or a general change of brightness in the image does not have a negative impact on the detection of defects. The reliability of the technique makes the defect detection completely unsusceptible to changes in illumination, such as brightness fluctuations or movement of the object.
A well-known method to expose surface defects on specular surfaces is with a stripe pattern. Using the Smartvision technology, all defects will be found; it is not necessary to create separate search areas for dark and light stripes.
BlobMax can be applied to objects made from ceramics, plastics, metal, paper, rubber or glass.
Currera Intelligent Vision System (IVS) is a compact, complete, easy to use and deploy PC-based imaging and vision solution, integrating camera and PC in single, matched and tuned industrial package.
The IVS platform is an open vision application and operating system, and fully supported line of Atom PC-based intelligent vision systems, overcoming the limitations of conventional industrial smart cameras.
Currera features integrated and well-matched imaging, interfacing and processing sub-systems, reducing the bandwidth bottlenecks as well as space, power and cabling requirements.
High performance fusion of imaging and processing hardware results in exceptionally low image acquisition requirements of less than 1 per cent of the CPU resources at 2.5Gbit/s acquisition bandwidth. This, and the fact the PC is an industry standard, offers a range of new design options and applications.
Sony Europe's Image Sensing Solutions (ISS) division has created one of the smallest C-mount camera modules in the industry. The device is on show and being launched at Vision 2010.
At 29 x 29 x 19mm the device, called Micro XCD, is nearly as small as a €2 coin. The module weighs just 37g and is 2/3 the cubic size of previous modules from Sony. The IEEE1394b I/F FireWire connected camera transmits data at 1.6MB/s and integrates a high quality CMOS image sensor with the latest FPGA technology; delivering high performance at low power consumption levels.
Sony has redesigned its PCB to incorporate thinner front and rear panels that are still very robust, making the camera suited to a wide range of industrial vision applications. Micro XCD devices will be available during Q4 2010.
Spotrack has developed a visual tracking system for tracking of a single object with multiple devices. The system is simple and versatile using a video camera to detect the object for tracking. Based on the object coordinates, the system calculates the correct amount of pan and tilt steps needed for each device to target the object.
Based on this method, a system for tracking stage performers with multiple lights (Spotrack) has been implemented; a multi-camera tracking system is currently under development. Devices for targeting (light, camera, etc) can be placed in an area with no restriction. Area dimensions are not needed, as the system doesn't use distance for pan and tilt step calculations.
The system can be implemented in a number of areas, such as in full automatic security systems (automatic detection and targeting) and robotics system among others.
Technos' system consists of one camera that can be positioned 1,800mm from an object to detect a 50µm defect. With this capability, the system can detect defects between pixels and therefore never miss a defect. The system can detect objects in 3D space. With one camera, the system can detect 1/115850 of visual field when the defect is shaped in points, and 1/1310000 when the defect is shaped in lines.
To detect an object 50µm with human eyes the human to be 200 to 240mm close to the object. A conventional CCD camera system needs to be positioned 75mm from the object.
Viz+ is a smart prototyping environment that allows users to test and try ideas quickly and conveniently. Users can edit the chain of processing steps using click and drag, and get instant feedback: operations are performed in real time and the effect of any parameter adjustment is immediately reflected. Watch points can be inserted anywhere and the evolution of the image processing and analysis steps can be seen, on still images or on live flow.
When the user is satisfied with the script, it can be exported as compiling source code. Viz+ wraps the Viz library, a set of general purpose tools for machine vision applications, which put emphasis on ease of use. Tools include image processing/analysis, pattern matching, gauging, OCR and code reading.
The Lynx is a high-performance shortwave-infrared imaging system based on Xenics' recently developed high-density line scan detector arrays. The GigE Vision-compatible Lynx camera is suited for scientific, machine vision, spectroscopy and optical coherence tomography (OCT) applications. Practically, all these applications require high line-rate sensors with high resolution and low noise combined with high dynamic range.
The Lynx camera is capable of providing all these features. It comes with either a 1,024 or a 2,048 pixel arrays at a line rate of 40kHz (10kHz for the 2,048 pixel array) and with an extremely small pixel pitch (12.5μm), which enables more accurate inspection and smaller particle detection at wavelengths where visual camera's (based on CMOS or CCD) have no sensitivity. It can be combined with a thermo-electric cooler (TEC) of up to three stages, offering low dark-current, and consequently high resolution combined with a high dynamic range.
Up until now, these resolution levels could only be obtained using complex multi-camera solutions. Use of the Lynx not only makes these types of setups simpler, but also cheaper.