Sensors/cameras shine at Vision 2007

Those looking for new sensors and cameras at Vision 2007 were not disappointed. Nearly every major manufacturer of machine-vision peripherals attended the show, and most introduced new products and technologies.

Jan 1st, 2008
Th 0801vsd Trends05

Those looking for new sensors and cameras at Vision 2007 were not disappointed. Nearly every major manufacturer of machine-vision peripherals attended the show, and most introduced new products and technologies. Many vendors also participated in “Industrial Vision Days,” a three-day seminar held on the show floor. Organized by the VDMA (Frankfurt, Germany), the seminar included more than 40 different presentations including selecting a lens for machine vision, the latest developments in the EMVA 1288 and GenICam standards, GigE vision cameras, and machine-vision software. Those who did not attend can download nearly all 40 of these presentations at www.vdma.org.

On the show floor, the introduction of new products was equally impressive. Image-sensor manufacturers Cypress Semiconductor (San Jose, CA, USA; www.cypress.com), Panavision Imaging (Homer, NY, USA; panavisionsvi.com), and Kodak (Rochester, NY, USA; www.kodak.com) chose the show to introduce three very different imagers. The Cypress LUPA-1300-2 features a 1280 × 1024 CMOS imager with 14 × 14-µm pixel size; its 12, 10-bit LVDS digital outputs running at 61-MHz pixel rate allow the imager to run at frame rates of up to 500 frames/s at full resolution.

While such imagers are firmly targeted at high-speed imaging, Panavision’s device, the Dynamix-35 XtremePX sensor, with a 5766 × 2164 pixel area and 8.5 × 2.8-µm pixels, seems more suited to HDTV applications. Like the LUPA, the Panavision device also features multiple LVDS outputs that allow it to run as fast as 120 frames/s. Kodak, a company firmly established in developing CCD imagers for machine-vision cameras, also showed its latest product, the KAI-01050, a 1/2-in. progressive-scan imager featuring 1024 × 1024 pixels.

“In designing a camera with the imager,” says Michael DeLuca, market segment manager for Kodak, “either one, two, or four output taps can be used. This allows the device to be clocked at 40 MHz and with a four-tap output to achieve 120 frames/s.” Such a scalable architecture is likely to extend to Kodak’s next generation of multiple-tap area-array cameras that may include a 2k × 2k device, although DeLuca was reluctant to comment on any future developments from the company. First to adopt the KAI-01050 was Adimec (Eindhoven, The Netherlands; www.adimec.com). At Vision 2007, the company showed its Opal 1000, a Camera Link camera running at 120 frames/s that incorporates the Kodak imager.

One of the most interesting demonstrations at the show was at the Dalsa (Waterloo, ON, Canada; www.dalsa.com) booth. For those interested in comparing the quality of CCD and CMOS imagers, the company displayed two of its cameras, one using a CCD imager, the other a CMOS imager, in a side-by-side comparison. While the Genie HM 1400 uses the company’s CMOS imager with 1400 × 1024 pixels, the Genie HM 640 uses the ICX424 from Sony (Park Ridge, NJ, USA; www.sony.com/videocameras). Although the HM 640’s imager is 640 × 480 pixels and the HM 1400 is 1400 × 1024 pixels, both imagers feature a 7.4-µm pixel pitch. In a visual comparison between the results of the same images captured with both cameras, there was not any discernable difference, indicating the strides Dalsa has made in the fabrication of its latest generation of CMOS imagers.

CMOS linescan cameras were also introduced by Basler (Ahrensburg, Germany; www.baslerweb.com) and e2v (Chelmsford, UK; www.e2v.com). Both companies demonstrated linescan cameras that use novel architectures to capture monochrome and color images at high speed. Basler’s latest Sprint Series of monochrome cameras, for example, uses a dual-line architecture in which the image sensor comprises either a 2k × 2 or 4k × 2 imager with no intermediate gap between the rows of photodiodes.


In time-delayed summing mode, a Basler linescan camera is triggered at up to 70 kHz (top). Each area of the object is scanned twice and time delay averaging used to reduce read noise and increase the camera’s signal-to-noise ratio. In dual-line mode, although the frame rate is 70 kHz, the effective line rate is 140 kHz because the camera captures two lines (bottom).
Click here to enlarge image

“The benefits of such a dual line sensors is that it can be used to double-scan the object being imaged,” says Lars Hansen, product manager. “Because each area of the object is scanned twice, time delay averaging reduces read noise and increases the camera’s signal-to-noise ratio by almost 3 dB. Such an architecture can be used to double the output line rate of the camera,” he adds (see figure). With a 312-kHz linescan rate (for the 4k × 2) camera, the Sprint Series features a Camera Link interface for data transfer to a host-based frame grabber.

“In the design of color linescan cameras,” says Christophe Robinet, strategic marketing manger of industrial products with e2v, “it is important to minimize the distance between the rows of RGB linear sensing elements to reduce any image artifacts that may occur when the sensor is used in high-speed applications.” In considering a color linescan camera for industrial systems, this spacing should be as low as possible.”

In the design of e2v’s EliiXA, a trilinear or quadrilinear CCD imager of 4k × 3 or 4k × 4 is used running at 18 kHz. “With 10-µm spacing between each row of monolithic CCD imagers, the ELiiXA features the lowest line spacing of any color linescan camera available,” claims Robinet. The ELiiXA cameras will be available in three versions in March 2008—one trilinear version (RGB) and two quad-linear versions: RGB-monochrome and RGB-NIR. Each camera will be available in full pixel rate (320 MHz) and in half pixel rate (160 MHz) with Camera Link interfaces.

More in Image sensors