Traditionally, machine vision components have been deployed by OEMs and sophisticated end users to build systems to improve the efficiency of manufacturing processes and to lower the cost and improve the quality of manufactured goods.
As the cost of the components used to build such systems continues to decrease, the advantage of deploying such systems in a manufacturing environment becomes ever more appealing as the price of the systems based on them become more affordable.
Over the past twenty years, many companies relocated their manufacturing facilities to the Far East to take advantage of lower labour costs. Today, however, as the demand for higher quality products increases and overseas labour costs rise, more manufacturing is returning to North America, and this is driving the need for new automated vision inspection systems.
Naturally enough, however, Far Eastern manufacturers keen to remain price competitive -- even with rising labour costs -- also see the advantage of employing such automated inspections systems, and the market growth for such systems may outstrip the growth in the North American and European market.
New markets, new opportunities
But the traditional markets for vision systems are not the only ones set to expand over the following decades. Outside the manufacturing arena, opportunities exist for vision suppliers in markets such as the life sciences, medical imaging, biometric analysis, intelligent traffic systems and surveillance fields. Even higher-volume opportunities exist in the embedded systems marketplace, where imaging systems are employed in PCs, smart phones, tablets and automobiles.
While these markets take advantage of many of the same hardware technologies that are deployed in the industrial arena, the more specialized fields have, and will continue, to give birth to a new set of software applications tailored to meet the needs of those applications.
Suppliers formulate new strategies
As the price of vision system components continues to fall, many suppliers who have solely offered single lines of vision components -- such as cameras -- will need to formulate strategies to ensure that they can continue to thrive in the marketplace. Typically, many such suppliers will do so by lowering their own manufacturing costs, improving their product quality and reducing their delivery times.
But there is also a trend for such component manufacturers to lower the total cost of ownership of the systems supplied to their customers. This will be achieved in one of two ways: through the in-house development of bundled software and hardware solutions, or through strategic joint relationships with manufacturers of complementary product lines. Both approaches will enable suppliers to provide more comprehensive offerings to their customers while alleviating interoperability issues. Similar agreements with software vendors will enable existing hardware manufacturers to deliver more complete system solutions.
Component suppliers with expertise in the imaging market may see also opportunities in diversifying into the production of sophisticated end user systems where profit margins may be higher. Having identified such opportunities in the emerging life sciences or surveillance fields, they too will need to form relationships with vendors, or individuals, outside the traditional machine vision market to bring such systems to fruition.
Higher density, faster speeds
Year upon year, the inexorable trend in the vision systems market has been toward cameras with higher density imagers, interfaces with faster transmission speeds, faster frame grabbers and the increased deployment of software that can take advantage of the processing power of multi-core processors and/or graphic processing units. This has inevitably led to the development of systems that can not only capture greater amounts of image data at faster rates, but process that data quicker, enabling system response times to be reduced.
While this trend looks set to continue into the future, certain vendors are taking advantage of new technologies with which they can distinguish themselves in the marketplace. In the field of lenses, for example, recent innovations include the development of liquid lenses that enable users to adjust the focal distance of a camera the over a specified range. In the camera arena, several vendors now offer cameras with multispectral capabilities, enabling their customers to capture images in the visible, NIR and/or SWIR wavelengths. In the lighting arena, standardized software interfaces are becoming available to enable OEMs to more easily configure and control the lighting in their systems. And on the software front, vendors are addressing the important issue of enhancing their software packages to enable systems integrators to interface the vision software to other control software in their systems.
From 2D to 3D
Traditionally, the majority of vision systems used in both the industrial and non-industrial marketplace have deployed 2D vision systems. While such systems may serve the majority of applications, a variety of 3D image processing hardware and software products are now becoming available, enabling OEMs to deliver systems with more sophisticated capabilities.
In the industrial marketplace, 3D vision systems are currently being deployed in a number of areas, where they can performs tasks such as the surface inspection of complex components. While such stand-alone vision systems are one area that 3D systems are making inroads, another area is in the field of robotics.
Figure 1: Kria’s T-EXSPEED vision system has been developed for speed enforcement applications.. The stereoscopic system monitors the road with two Prosilica GT cameras from Allied Vision Technologies. Using a series of images, the software can determine which distance the vehicle covered in the time interval between two images and calculate its exact speed up to 300 km/h. It also detects single and multiple red light violations.
Here, 3D vision systems can provide robots with the intelligence to perform roles such as depalletizing, box moving and random bin picking in factory environments. Outside the traditional machine vision marketplace, 3D systems are also making inroads into a variety of disparate areas, in the transportation, medical life sciences and agricultural fields.
A transformation in vision software
While changes in the hardware used in vision systems over the next ten years might be incremental in nature, the same cannot be said for the software that will be used by OEMs to develop those vision systems. The past ten years has seen a transformation in the vision software industry, with many vendors now offering drag and drop software programming environments that can be deployed by their customers to develop vision systems quickly and effectively with little or no former programming knowledge.
Figure 2: At the London Container Terminal in Tilbury, Essex, UK, Visy Oy has deployed a system based on Basler IP cameras that capture images of every vehicle approaching the gate barriers at the facility. The images are then transmitted to an access control operating room, where OCR detection takes place. If the system associates the detected license plate number with an authorized vehicle, the barrier will open and let the vehicle pass.
While future vision systems will still need to capture images, process them and effect an operation based on that image analysis, whether the image capture is still performed by CMOS or CCD sensors, or whether those sensors are still interfaced to a traditional set of hardware, will become, in a sense, immaterial. The most important consideration may well be the sophistication, and ease of use, of the software that will enable engineers to configure such systems and to ensure that they operate effectively in industrial, or non-industrial, environments.
Today’s software may be effective, however, software vendors now realise the importance of enhancing their offerings to enable their vision packages to interface with system software used to control or monitor other aspects of systems such as sensors, electric actuators, and motors.
In the future, systems integrators may be able to develop automated vision inspection systems by plugging together a set of virtual hardware over a set of standard virtual interfaces. Once that has been achieved, advanced system software might automatically identify the capabilities of the virtual hardware in the system and configure all of the parameters of the inspection task through a single user interface. Hence, systems integrators may ultimately be able to run a virtual inspection of a product even before any hardware has been purchased.
Figure 3: Universal Robotics has developed a 3D system that uses four inexpensive Microsoft Kinect sensors mounted above and below a conveyor to capture 3D images of pallets. The images are then analyzed for defects such as missing wood, notches or damage. The system is claimed to be 300%faster than the previous manual inspection process.
Needless to say, the development effort required to produce such systems may be a long way into the future! But should such systems become available, it may not simply change the role of the traditional systems integrator but may enable more sophisticated end users to take on the task of developing such automated inspection systems themselves.
Written by Dave Wilson, Senior Editor, Novus Light Technologies Today
This article is a part of Novus Light Technologies Today's exclusive series celebrating The International Year of Light 2015.