Today, many camera vendors develop products based around standard sensors from manufacturers such as Aptina (San Jose, CA, USA; www.aptina.com), Sony (Tokyo, Japan; www.sony.com), and Eastman Kodak (Rochester, NY, USA; www.kodak.com/go/imagers). Because of this, it has become increasingly difficult for camera companies to differentiate their products from their competitors’.
Realizing this, many camera manufacturers are now teaming up with CMOS imager manufacturers, often aiding these vendors in the specifications of new devices. Nowhere was this more apparent than at VISION 2009 in Stuttgart where a number of new CMOS imagers and cameras were displayed.
Although some camera vendors did not wish the fruits of these alliances to be made public, Point Grey Research (Richmond, BC, Canada; www.ptgrey.com) and Imaging Development Systems (IDS; Obersulm, Germany; www.ids-imaging.de) both announced cameras they had developed in conjunction with CMOS imager vendors.
For its part, Point Grey Research teamed up with CMOSIS (Antwerp, Belgium; www.cmosis.com), embedding the company’s 1-in., 5.5 × 5.5-μm, 2048 × 2048-pixel, 170-frames/s CMV4000 and 2/3-in., 5.5 × 5.5-μm, 2048 × 1088-pixel, 340-frames/s CMV2000 into its line of Gazelle Camera Link cameras.
While high-resolution, high-speed cameras were the theme of Point Grey’s product introductions, IDS used the MD1-10-B Logarithmic sensor from New Imaging Technologies (NIT; Evry, France; www.new-imaging-technologies.com) in the design of its latest high-dynamic-range (HDR) uEye camera. With a dynamic range of 120 dB, the camera features a 768 × 576-pixel imager in a 1/1.8-in. format, operates at 50 frames/s, and will be offered in USB and GigE versions.
Interestingly, many different sensor architectures have been used to increase the dynamic range in both CCD and CMOS sensors (see “Dynamic Design,” Vision Systems Design, October 2009). However, until now, most of these devices have used architectures such as dynamic well capacity adjustment, multiple image capture, or pseudo-logarithmic transfer functions to achieve this. Rather than incorporate such methods, Yang Ni, president of NIT, developed a true logarithmic sensor using a photodiode operated in photovoltaic mode.
“In most CMOS logarithmic designs,” says Ni, “linear photo current passes through a sub-threshold operating MOS transistor and the exponential characteristic of this transistor transforms the linear photo current into a logarithmic-like voltage output signal.”
However, such designs result in a large amount of fixed pattern noise (FPN) resulting from dark signal and photo response nonuniformity. To overcome this problem, the Matrice Active à Génération d’Image indexée sur Contraste (MAGIC) technology developed at NIT uses a pixel design based on a photodiode operated in photovoltaic mode. In this design, the image signal is formed by the open-circuit voltage of a photodiode under illumination (see figure).
In the MAGIC pixel design, the photodiode is operated in photovoltaic mode with the result that the open-circuit voltage across the pn junction is proportional to a logarithmic value of the incident light intensity.
“Because this photovoltaic mode photodiode is read by a DC-isolated amplifier, a large amount of FPN is also introduced and must be reduced. This is accomplished by connecting a MOS transistor across the photovoltaic photodiode,” says Ni. When activated, this MOS transistor zeros the voltage of the photodiode under illumination and gives an exact dark reference voltage.
By using on-chip readout, the FPN caused by the in-pixel buffer amplifier can then be removed. Better still, because this on-chip FPN correction operation resets the photodiode into a zero bias state before each image is captured, the image lag, inherent in a conventional logarithmic photoreceptor, is eliminated.
IDS used the NIT imager for machine-vision camera applications. NIT also provides the MCD1-10-BW, an analog camera module that uses the HDR imager. According to Ni, NIT will focus on other markets such as security, automotive, and consumer applications using either this image sensor or incorporating this HDR technology into custom-designed smart image sensors.
