How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
While many image processing products and systems are offered as being high-speed, there is no generally accepted definition for high speed. There are two issues: one is how speed is measured, and the other is what criteria determine that a system is high speed.
One way to measure image processing speed is the time between the trigger signal and the results being output from the image analysis. As a benchmark, depending on the size of the image and the processing performed, it is possible to achieve times of less than 10 milliseconds.
Another way to measure speed is the rate at which a system can accept input images. If the acquisition rate is high, then the system is considered high-speed even if the time to provide an output is longer than the time between input images. It is common to use image buffers and cascaded processors to design a system that accepts images at a high rate with the time between start of acquisition and the final results being longer than the time between input images. For this measure of speed, the only limitation is the amount of processing resources employed. Using components specifically designed for this purpose, it is possible to achieve sustained rates of several hundred images per second (see Figure 1.)
The criteria for determining when a system is high-speed can be either associative or comparative. If an image processing system is used in an environment that itself is considered high-speed, and the image processing system can perform at the environment’s pace, the, by association, the image processing system is high-speed.
With comparative criteria, if a certain image processing system or component is able to input or process images faster than most other similar products, then in comparison to those other products, it is high-speed.