How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Cameras with Gigabit Ethernet (GigE) interfaces are cropping up in imaging applications everywhere, and there is good reason for that. GigE delivers full-duplex connectivity at one gigabit per second (Gb/s) over standard, low-cost infrastructure. It has long-distance reach, supports almost any network configuration that can be envisioned, and allows heavy-duty processing tasks to be handled by scaleable stacks of economical PCs.
All GigE cameras look about the same from the back, with a RJ-45 LAN plug. But when it comes to performance, the interface electronics inside can make a world of difference.
GigE camera interfaces must perform, at minimum, six functions:
Basically, two design approaches can be taken to meet these requirements: build purpose-built hardware from the ground up, or write a software application for an embedded processor. Deciding which approach to take can be tricky and, in the end, usually reflects a combination of application needs and business priorities.
Hardware = Low Power, High Performance
In general, purpose-built hardware yields a high-performance, reliable GigE interface that can accommodate just about any camera sensor, including those with high resolutions and fast frame rates. Hardware interfaces operate with clock-cycle accuracy, and so perform processing tasks quickly, efficiently, and deterministically. They fit into compact footprints, important for small-body cameras.
Hardware interfaces also draw a small amount of power — as low as 2.0 watts (W) — and this level varies only slightly with different sensors. This is a key point, and one that gives hardware interfaces a clear advantage over software solutions, because it means camera manufacturers can use one interface to GigE-enable a wide range of cameras in their portfolios.