How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Machine vision integrators are facing ever-increasing pressure to extract large amounts of information about the object of interest at high speeds. Among several approaches adopted by the industry, systems integrating large format cameras with a double-telecentric lens seem to be gaining in popularity, particularly in metrology and gauging applications.
The semiconductor industry continually has been developing sensors with smaller and smaller pixels in the interest of increasing resolution. However, there are limitations to the pixel size and acceptable signal-to-noise ratios. Using a large-format sensor with reasonable pixel size is the preferred approach when needing to sample large volumes of information quickly and accurately. In order to accommodate the majority of the large-format sensors currently on the market, the imaging lens needs to cover an image diagonal dimension of at least 30 mm and up to as much as 90 mm. Also, to ensure sufficient contrast, the toleranced Modulation Transfer Function (MTF) of the lens should be better than 0.20 in the resolution range of 70 lp/mm to 100 lp/mm.
A telecentric lens is typically selected to minimize parallax (perspective) errors in the image. By definition, the entrance pupil of a telecentric lens is located at infinity, which makes them ideal for reducing dimension-measurement errors caused by focus adjustments and test-site vibration. High-quality telecentric imaging lenses offer uniform close-to-diffraction-limited image resolution performance over the entire field of view. The lenses also typically have less than 0.1 percent distortion and no vignetting across the whole image plane.
A double-telecentric lens has both the entrance and exit pupils located at infinity. These lenses actually are an afocal optical system working as a finite-conjugate imaging lens. In an afocal system the image magnification for near field objects is theoretically constant for all object distances. This magnification stability eliminates magnification variance due to lens/camera manufacturing errors.
Double-telecentric lenses also offer several advantages in image formation. A non-vignetted lens enables each pixel to achieve optimal sensitivity since the chief rays are normal to the sensor plane. Another key benefit is the reduction of off-axis irradiance fall-off caused by the cosine-fourth law. This is extremely important for all large-format imaging applications. For example, a double-telecentric lens with a maximum sensor diagonal of 30 mm designed by Edmund OpticsŪ achieves an off-axis relative illumination of 86 percent.