On-the-fly defective pixel correction
AssigneeSRI International
InventorsRui Zhu, Peter A Levine, Colin Earle
Life of this patent
Estimating the term…
Abstract
An image sensor with a pixel array that has two or more pixels is discussed. The defective pixel detector detects a defective pixel. A dynamic threshold range is based on amplitude levels of neighboring pixels. The defective pixel detector detects the defective pixel with the dynamic threshold range to detect and determine a defective status for the pixel under analysis based on an amplitude level of the pixel under analysis relative to the amplitude levels of neighboring pixels. The defective pixel detector applies algorithms to create candidate pixel patterns for the neighboring pixels used to create the dynamic threshold range. The value of the dynamic threshold range will change based on the amplitude levels of the neighboring pixels.
Claims17
An apparatus, comprising:
- an image sensor with a pixel array, where the pixel array has two or more pixels,
- a defective pixel detector that is configured to detect a defective pixel, where a dynamic threshold range is based on amplitude levels of neighboring pixels,
- where the defective pixel detector is configured to detect the defective pixel with the dynamic threshold range to detect and determine a defective status for the pixel under analysis based on an amplitude level of the pixel under analysis relative to the amplitude levels of neighboring pixels, and
- where the defective pixel detector is configured to apply one or more algorithms to create candidate pixel patterns for the neighboring pixels used to create the dynamic threshold range, where a value of the dynamic threshold range will change based on the amplitude levels of the neighboring pixels, where an amplitude detector is configured to determine the amplitude level of the pixel under analysis and then supply that to the defective pixel detector, where the amplitude level detected by the amplitude detector includes at least a dark current noise determined by one of i) a measurement of the dark current noise or ii) a mathematical determination of the dark current noise of the pixel under analysis.
A method for an image sensor, comprising:
- configuring an image sensor with a pixel array, where the pixel array has two or more pixels,
- configuring a defective pixel detector to detect a defective pixel, where a dynamic threshold range is based on amplitude levels of neighboring pixels,
- configuring the defective pixel detector to detect the defective pixel with the dynamic threshold range to detect and determine a defective status for the pixel under analysis based on an amplitude level of the pixel under analysis relative to the amplitude levels of neighboring pixels,
- configuring the defective pixel detector to apply one or more algorithms to create candidate pixel patterns for the neighboring pixels used to create the dynamic threshold range, where a value of the dynamic threshold range will change based on the amplitude levels of the neighboring pixels, and
- configuring a substitute pixel value generator to generate a pixel correction value when a comparison is made between i) a pixel value of the pixel under analysis to ii) a dynamic threshold range plus a fractional factor, and when the pixel value of the pixel under analysis is outside the dynamic threshold range plus the fractional factor, then the substitute pixel value generator is configured to i) determine that the pixel under analysis is defective, and then ii) substitute 1) the pixel correction value for 2) the pixel value of the defective pixel.
A system, comprising:
- a camera with an image sensor with a pixel array, where the pixel array has two or more pixels,
- a defective pixel detector configured to detect a defective pixel, where a dynamic threshold range is based on amplitude levels of neighboring pixels,
- where the defective pixel detector is configured to detect the defective pixel with the dynamic threshold range to detect and determine a defective status for the pixel under analysis based on an amplitude level of the pixel under analysis relative to the amplitude levels of neighboring pixels,
- where the defective pixel detector is configured to apply one or more algorithms to create candidate pixel patterns for the neighboring pixels used to create the dynamic threshold range, where a value of the dynamic threshold range will change based on the amplitude levels of the neighboring pixels,
- a substitute pixel value generator configured to generate a pixel correction value when the defective pixel detector determine the defective status for the pixel under analysis,
- a random telegraph noise detector, a dark current noise detector, and a pixel read noise detector are configured to detect the noise level of the neighboring pixels, where the random telegraph noise detector, the dark current noise detector, and the pixel read noise detector are further configured to combine to provide a noise input into the defective pixel detector, and
- where the defective pixel detector is further configured to couple to and communicate with a substitute pixel value generator in order to generate a pixel correction value for the defective pixel based on a dynamic threshold range determined from inputs of the random telegraph noise detector, the dark current noise detector, the pixel read noise detector.
Dependencies are inferred from the claim text.