Viewpoints
@imveurope
www.imveurope.com
processing algorithm, application and system requirements – power, temperature range, etc. – enable an end user to identify the most capable embedded hardware platform. And, when it comes to embedded processors, there are plenty to choose from with thousands of boards available on the market today. Where an Arm chip requires less power,
GPUs – which may be the most underutilised processors available – offer a host of powerful capabilities, though they’re still not suited to every image processing algorithm. In fact, not all image processing should be embedded. In some cases, CPUs offer the fundamental architecture needed to support a customer’s application algorithm more effectively than GPU processors. Customers must know their own design and understand their own image processing algorithm to know which processor, embedded or not, should be used. In making these decisions, customers also
Manny Romero, product manager at Teledyne Dalsa, says embedded processing will reduce costs and improve efficiency for machine vision customers in 2016 and beyond
V
alue – the functionality and performance received for the price paid – will continue to remain a
primary factor in decision making for machine vision customers for the remainder of 2015, as well as into 2016 and beyond, and the shiſt toward embedded processing closely aligns with the desire of customers to realise greater value. Simply put, customers don’t want to pay for features and components they don’t need, and embedded processors can help customers maximise their investments in machine vision solutions. By taking advantage of
type processor, which offers low power consumption, lower cost and greater efficiency align as lower power requirements make operations more efficient overall. An interesting side benefit of embedded
GPUs offer a host of powerful capabilities, though they’re still not suited to every image processing algorithm
embedded processing, customers are able to reduce the total number of system components required, designing a solution customised to meet the specific goals of their application and lowering the total cost. At the soſtware level, for example, there is a
push toward less power hungry and less costly operating systems like Linux, which can help ensure that the power used by the system is dedicated toward application performance rather than system operation. When this type of operating system is combined with an embedded processor such as an Arm-
processing is that in addition to enabling more efficient, cost-effective machine vision solutions, embedded processors can also enhance an application’s flexibility. With lower power consumption, these applications are able to sustain greater variations in temperature, making them suitable for harsher environments and expanding the number of places where
they can be deployed. Plus, fewer components mean a smaller total footprint, making machine vision solutions with embedded processors adaptable to an even broader range of application environments.
Realising the full potential of embedded architectures Te most critical consideration for those seeking to get the greatest value from an embedded architecture takes place during the design of the machine vision application, when a clear understanding of the image
14 Imaging and Machine Vision Europe • Yearbook 2015/2016
should understand whether the vendors they’re considering offer products that are truly interoperable with the machine vision solution they’ve designed. For example, will the vendor’s soſtware allow its products to operate on the platform that best supports the customer’s application, and then, depending on interface requirements, will the product support interfaces to embedded systems in GigE, USB, CSI-2 (MIPI) or PCI-e, as required? For an application to operate efficiently – and for a customer to realise the value they expect – it is vital to ensure that all the right components are in place.
What’s next for embedded processing? With continuing requirements for improved cost-efficiencies, more and more processing will be performed ‘on board’ in 2016, and the move toward embedded processors will only continue to grow in the years to come. However, even with a focus on the expanded use of embedded processors, it is also essential to remember the linkages between all elements of an effective machine vision solution, whether processors, sensors, cameras, or lenses. Customers will seek those offerings that promise measurable value: high-performing, highly efficient systems delivered at the right price. In response, vendors will continue to
innovate so that they can deliver the cost- effective machine vision solutions their customers demand, and retain their positions in a highly competitive marketplace.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48