Charge Injection Device (CID) cameras have been in use since the early 1970's, but it has only been in the last several years that knowledge and experience have combined for a broader understanding of the technology and how to more fully apply it.
The CID concept was originated by scientists at General Electric Company working to devise a semiconductor memory chip. Exploiting the photosensitive characteristics of silicon, they developed a simple X,Y addressable array of photosensitive capacitor elements, and evolved the first CID camera in 1972. Development efforts continued through the 70's and 80's resulting in the generation of some 30 patents describing the basic structure and readout techniques presently employed. CIDTEC was established when management took the business private through a leveraged buyout in July of 1987.
Every pixel in a CID array can be individually addressed via electrical indexing of row and column electrodes. Unlike Charge Coupled Device (CCD) cameras which transfer collected charge out of the pixel during readout (and hence erase the image stored on the sensor), charge does not transfer from site to site in the CID array. Instead, a displacement current proportional to the stored signal charge is read when charge "packets" are shifted between capacitors within individually selected pixels. The displacement current is amplified, converted to a voltage, and fed to the outside world as part of a composite video signal or digitized signal. Readout is non-destructive because the charge remains intact in the pixel after the signal level has been determined. To clear the array for new frame integration, the row and column electrodes in each pixel are momentarily switched to ground releasing, or "injecting" the charge into the substrate.
This principle of operation makes CID technology fundamentally different from other imaging techniques, giving rise to a number of technical advantages that can be used to solve imaging problems. For instance, the nondestructive readout capability of CID cameras makes it possible to introduce a high degree of exposure control to low-light viewing of static scenes. By suspending the charge injection, the user initiates "multiple-frame integration" (time-lapse exposure) and can view the image until the optimum exposure develops. Integration may proceed for milliseconds or up to several hours with the addition of sensor cooling, applied to retard accumulation of thermally-generated dark current. Controlled integration is useful for scientific and photographic applications, especially in astronomy.
At brighter intensities, "blooming" and "smearing" describe the
distortion in an image that can occur when solid-state video cameras
are exposed to concentrated, non-uniform light as in the images
Charge can spill from oversaturated elements to adjoining pixels, or the shift registers (charge transfer mechanisms) during readout, eradicating portions of the image. By contrast, CID imagers are more tolerant to intense light because optical overloads are confined to illuminated pixels. Charge is not exported from the pixel collectors, so the structure offers no paths along which overloads can propagate, and radial spreading of charge is minimized as the excess is drawn into the underlying charge collector. This inherent antiblooming performance ensures accurate image detail even under extreme lighting conditions, so CID cameras have been used effectively for missile tracking, semiconductor pattern recognition, and factory inspection where reflections and the appearance of bright objects give rise to brilliant light intensities within a properly exposed image.
The contiguous pixel structure of CID arrays further contributes to accurate imaging because there are virtually no opaque areas between pixels where image detail can be lost. This attribute is important for applications where precise dimensional data is critical, particularly in the determination of object edges for inspection, measurement, positioning, and tracking. Employing interpixel processing techniques, CID cameras are currently utilized in precision gauging equipment performing measurements accurately to half a micron.
The uniform topological structure of the imagers provides homogeneous pixel-to-pixel response to coherent illumination for more accurate reproduction of laser profiles; ideal for beam diagnostics and analysis. CID sensors also offer wide spectral response, from 200 to 1100 nanometers, allowing capture of images produced by light sources ranging from UV to the near IR. And the PMOS structure reduces the effect of radiation on sensor operation, making CID's less vulnerable to disruption in low-level radiation environments than NMOS devices (structure used in many CCD's). Radiation hardened CID's are currently employed in nuclear power, industrial X-ray, scientific, and space applications. They are also applied in several classified military programs.
Since each pixel in the CID array can be addressed individually, flexible readout and processing options are made possible. For example, "Progressive Scan" readout enables real-time processing by eliminating the delay required to combine odd and even fields (2:1 Interlace scanning). Instead, lines are read sequentially (1, 2, 3, 4, etc.) allowing an image processor to analyze the latest row of video information while readout continues to the next line. The 60 frame per second output of these cameras provides high-speed operation without sacrificing RS-170 compatibility, so they interface with RS-170 frame-buffers, TV monitors, and VCR's.
For more efficient computer interfacing, binary format CID arrays (512 X 512, 256 X256, 128 X 128) are available to match standard memory formats. Their square pixels simplify computational algorithms, reducing processing complexity. Cameras incorporating these arrays are designed with advanced features that maximize image processing capabilities, so some models do not adhere to the limiting RS-170 timing standards established for broadcast TV. However, several different vendors now offer interface boards that integrate these cameras with RS-170 system components such as TV monitors and recording equipment.
Progressive scanning also opens the door to a number of CID camera features that expand the range of user options. For instance, in applications where full frame resolution is not required, but faster capture is, "Frame Reset" allows camera users to decrease vertical frame size for higher frame rates. With fewer lines to read, "shorter" frames can be read out faster. Frame Reset ends frames under user control by resetting the camera for new frame scanning.
For applications where only small regions in the field of view require attention at any given time, the "Rapid Scan" feature allows the user to isolate multiple areas of interest, or "windows" for readout at normal rates while scanning at a very high rate between windows. The selective data extraction speeds readout and reduces data volume, facilitating high-speed processing. This capability is especially powerful for target acquisition where several objects must be tracked individually, at rates up to several hundred images per second as they move through the field of view. Windowing is also used for high-speed inspection where specific sections of a pharmaceutical bottle, for example, are read out for processing to check cap placement or verify expiration dates and lot codes.
Available Random Access CID arrays (RACID's) further extend user control by providing the capability to address specific pixels in any sequence for selective readout at maximum scanning rates. Readout sequence is software controlled. RACID's are successfully employed in star tracking and celestial navigation applications where only designated stars within the canopy of space are read and processed for positional vectoring.
The stop motion or "freeze frame" capability of CID cameras enables them to accurately capture and read asynchronous high-speed events. CID operation allows image capture to proceed independent of camera timing, so the user times the camera to the event instead of timing the event to the camera's "vertical blanking interval" (period between frames when scanning returns to the top of the array in preparation for new frame readout).
As the bottle travels down the production line, a vision system senses when it has moved into position within the camera's field of view and engages the Inject Inhibit feature. Sensor scanning continues, but readout is suspended allowing integration to continue uninterrupted. The vision system fires a strobe at the appropriate moment and the image of the bottle is captured on the sensor. At the onset of the vertical blanking interval, Inject Inhibit is released returning the camera to normal operation. Readout of the captured image proceeds at the completion of blanking as new frame scanning begins. Asynchronous image capture is achieved by utilizing the Inject Inhibit function to "hold" the image momentarily until scanning of the next full frame begins, allowing readout of the complete image.
The Frame Reset feature increases throughput speed in freeze frame applications. When the bottle in our example moves into position, the vision system triggers Frame Reset, which resets the camera for new frame scanning. The strobe is fired and the image is captured on the sensor during the vertical blanking interval. Image readout proceeds at the completion of blanking when new frame scanning begins. The vision system resets the camera asynchronously in response to random events, providing almost immediate readout of the captured images. 60 FPS CID cameras asynchronously capture and read out well over 1800 images per minute. Implementing windowing or reading smaller frames allows even higher throughput rates.
CID overview provided by: CID Technology, 101 Commerce Blvd, Liverpool, NY 13088 (315) 451-9421