Pixel correction system and method for CMOS imagers

Details Pixel correction system and method for CMOS imagers Pixel correction system and method for CMOS imagers

1997-11-05

2003-09-09

Vu, Ngoc-Yen (Department: 2612)

Television

Camera, system and detail

Combined image signal generator and general image signal...

C348S310000

Reexamination Certificate

active

06618084

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates to an integrated circuit architecture for CMOS imagers. More specifically, the present invention relates to methods and apparatus for masking or correcting faults in individual pixels of CMOS imagers.
CMOS image sensors are now becoming competitive with charge coupled device (“CCD”) array image sensors. Potential applications include digital cameras, night time driving displays for automobiles, and computer peripherals for document capture and visual communications.
Since the 1970s, CCD arrays have dominated the electronic image sensor market. They have outperformed CMOS array sensors in most areas, including quantum efficiency, optical fill factor (the fraction of a pixel used for detection), charge transfer efficiency, readout rate, readout noise, and dynamic range. However, the steady improvement in CMOS technology (including increasingly small device size) has moved CMOS image sensors into a competitive posture. Further, in comparison to CCD technology, CMOS technology provides lower power consumption, increased functionality, and potentially lower cost. Researchers now envision single chip CMOS cameras having (a) integrated timing and control electronics, (b) a sensor array, (c) signal processing electronics, (d) an analog-to-digital converter, and (e) interface electronics. See Fossum, “CMOS Image Sensors: Electronic Camera On A Chip,” 1995 IEDM Technical Digest, Wash. D.C., Dec. 10-13, 1995, pp. 17-25 which is incorporated herein by reference for all purposes.
CCD arrays are limited in that all image data is read by shifting analog charge packets from the CCD array interior to the periphery in a pixel-by-pixel manner. Unfortunately, the pixels of the CCD array are not randomly addressable. In addition, due to voltage, capacitance, and process constraints, CCD arrays are not well suited to integration at the level possible in CMOS integrated circuits. Hence, any supplemental processing circuitry required for CCD sensors (e.g., memory for storing information related to the sensor) must generally be provided on separate chips. This, of course, increases the system's cost.
A persistent problem of both CMOS and CCD image sensor technologies is image degradation due to faulty pixels. Such faulty pixels arise from processing variations inherent in fabrication of numerous sensor chips. A pixel's fault may be manifested by an output indicative of a radiation exposure that does not accurately reflect the actual radiation exposure to which the pixel was exposed. For example, a pixel that outputs more charge than is expected upon exposure to a particular amount of radiation appears as a bright spot in an image. Similarly, a pixel that outputs less charge than expected appears as a dark spot.
Typically, image sensors are tested after fabrication to identify the number of faulty pixels that they contain. If any sensor has more than a specified number of faulty pixels, it must be rejected. Thus, sensor yield is limited by the number of faulty pixels typically produced on a chip. Not surprisingly, wide area sensors having large numbers of pixels have relatively low yields because they tend to have higher numbers of faulty pixels (the number of faulty pixels per total number of pixels is approximately constant for a given fabrication technology).
While careful screening of image sensors after fabrication can locate defective arrays, it cannot prevent sensors from degrading during normal use. Pixels sometimes develop such faults during normal operation. Unfortunately, no effective mechanism exists for identifying and correcting such faults.
What is needed therefore is an improved image sensor that can mask or otherwise correct defective pixels soon after the sensor is fabricated and during its lifetime.
SUMMARY OF THE INVENTION
The present invention provides a fault tolerant radiation imager such as a CMOS imager. Such image sensor includes circuitry for masking and/or correcting defective pixels during image generation. Masking may involve, in one example, replacing the output of a given pixel with an average of the outputs of surrounding non-faulty pixels. Correction may involve increasing or decreasing the output level from a pixel to be corrected. Thus, while image sensors may be fabricated with some number of faulty pixels, the images produced by such sensors will be superior to those produced by sensors that do not have such masking or correcting mechanisms. Further, sensor yield.may be improved because even those sensors having more than the traditionally allotted number of defects may be used (so long as the defective pixels are appropriately masked or corrected). In one preferred embodiment, pixels that were initially good and later become defective may be identified at the time they become defective. The newly defective pixels so identified may then be masked or corrected as described to thereby increase the CMOS detector lifetime.
One aspect of the present invention provides a method of masking faulty pixels that may be characterized by the following steps: (a) exposing an imager to radiation; (b) identifying a faulty pixel in the imager; (c) determining outputs of a plurality of other pixels located about the faulty pixel; and (d) masking an output of the faulty pixel using the outputs of the plurality of other pixels. Then a masked output of the faulty pixel can be provided in an image produced by the imager. Preferably, the system identifies the faulty pixel by accessing memory associated with the imager to determine the location of the faulty pixel in the imager. Then to generate an image corrected for the faulty pixel, the method may also require conversion of the output of the pixels (including the masked output) from an analog format to a digital format prior to masking the pixel's output.
In one embodiment, masking the output of a faulty pixel involves replacing its output with an output value based exclusively on the output of multiple surrounding pixels. For example, the corrected output may be obtained by interpolating output values of at least two pixels surrounding the at least one faulty pixel. In color imagers, the pixels of the imager can discriminate between multiple colors and the faulty pixel therefore will be designated to detect radiation of a single color. Preferably, in such cases, the plurality of other pixels chosen to mask the output of the faulty pixel are only those surrounding pixels that are designated to detect radiation of the same color as the faulty pixel.
It may not be necessary to totally mask the output of each of the faulty pixels. For example, some pixels may not be completely corrupted, in which case their outputs could be corrected rather than masked. In one instance, the faulty pixel may be known to have an output that is 20% lower than appropriate. Thus, the output of that pixel may simply be increased by 25% before producing the image.
Of course, some technique for identifying faulty pixels must be conducted initially (when the imager is produced) or periodically over its lifetime. In one embodiment, the system electronically tests a selected pixel according to the following sequence: (i) electronically resetting the selected pixel to a defined charge (e.g., the charge associated with a dark value in an image); (ii) reading the selected pixel's output; and (iii) comparing the selected pixel's output to an expected value based upon the defined charge provided to the selected pixel. If the selected pixel's output significantly deviates from the expected value, the system designates the selected pixel as faulty. Alternatively, the conducted test may also be an optical test in which the imager is exposed to radiation of a known intensity for a known duration. The outputs of the pixels are then compared to expected values as described above.
In the case of an electronic test and where the selected pixel includes an n-well and a p-diffusion photodiode, electronically resetting the selected pixel requires injecting a quantity of negative charge into the

Affiliated with

Also associated with

Rating

Pixel correction system and method for CMOS imagers does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Pixel correction system and method for CMOS imagers, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Pixel correction system and method for CMOS imagers will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3005972