Apparatus and method for compensating for fixed pattern noise in planar arrays

Claim 2

Apparatus for reducing fixed pattern noise in an image observed by an array of detectors, said apparatus comprising a plurality of image-responsive detectors, each said detector producing an image signal responsive to the portion of an observed image incident on said detector, thereby creating a set of image signals, dithering means for scanning said observed image across said detectors in a dithering cycle, temporal high-pass filtering means having a non-zero DC response for generating a set of filtered image signals each of which corresponds to one of said image signals, each of said filtered image signals having an attenuated low frequency component, and image restoration means for generating a reconstructed image based on said set of filtered image signals, wherein said reconstructed image corresponds to said observed image having reduced fixed pattern noise.

Claim 3

Apparatus according to claim 1 wherein said plurality of image-responsive detectors is a two-dimensional array of infrared detectors.

Claim 4

Apparatus according to claim 1 wherein said dithering means comprises means for repeatedly moving said detectors relative to the observed image.

Claim 5

Apparatus according to claim 3, wherein said means for moving said detectors is a motorized gimbal for changing the angular position of said detectors.

Claim 6

Apparatus according to claim 1 wherein said dithering means comprises means for repeatedly moving the observed image relative to said detectors.

Claim 7

Apparatus according to claim 5, wherein said means for moving the observed image is a steerable mirror in the optical path of said detectors.

Claim 8

Apparatus according to claim 6, wherein said steerable mirror is a galvo mirror.

Claim 9

Apparatus according to claim 6, wherein said steerable mirror is a piezoelectric actuated mirror.

Claim 10

Apparatus according to claim 1, wherein said dithering means includes a base supporting said plurality of detectors, said base having an inherent jitter motion.

Claim 11

Apparatus according to claim 1, wherein said temporal high-pass filtering means includes a finite impulse response filter.

Claim 12

Apparatus according to claim 1, wherein said temporal high-pass filtering means includes an infinite impulse response filter.

Claim 13

Apparatus according to claim 1, further comprising a shifting means for periodically sampling each of said image signals at a period corresponding to an integer multiple of said dithering cycle, said shifting means forming an output signal representing the observed image with reduced fixed pattern noise.

Claim 14

Apparatus according to claim 3 wherein said image, filtered image restoration means further includes convolving means for convolving said signals with an inverse kernel of said temporal high-pass filtering means.

Claim 15

Apparatus according to claim 1, further comprising correcting means for iteratively determining the fixed pattern noise in said observed image as a function of said image signal, said reconstructed image, and an old noise signal representative of a previously determined fixed pattern noise in a previously observed image.

Claim 16

Apparatus according to claim 14, wherein said correcting means comprises differencing means for generating an offset update signal corresponding to the difference between said reconstructed image and said image signal, and updating means for generating a noise correction signal corresponding to the sum of said offset update signal and said old noise signal.

Claim 17

Apparatus according to claim 15, wherein said updating means includes memory means for storing and retrieving said old noise signal, and an adder coupled with said memory means for calculating the sum of said offset update signal and said old noise signal.

Claim 18

Apparatus according to claim 15, wherein said differencing means further includes an offset estimate gain circuit for modifying said offset update signal.

Claim 19

Apparatus according to claim 17, wherein said offset estimate gain circuit varies over time to optimize reduction of fixed pattern noise.

Claim 20

Apparatus according to claim 17, wherein said offset estimate gain circuit is a scalar factor.

Claim 21

Apparatus according to claim 15, further comprising voter means for reducing error in said noise correction signal.

Claim 22

Apparatus according to claim 1, further comprising frame averaging means for reducing temporal noise.

Claim 23

Apparatus according to claim 1, further comprising an object acquisition means for generating an acquisition signal that identifies the presence of said observed image.

Claim 24

Apparatus according to claim 22, including a temporal high-pass filter for generating a first set of signals having an attenuated low frequency component, said first set of signals corresponding to said image signals, a sampler for periodically sampling said first set of signals at a period corresponding to an integer multiple of said dithering cycle, and an averager for generating said acquisition signal, said acquisition signal corresponding to the average of said sampled first set of signals.

Claim 27

A method according to claim 24, wherein said dithering step includes repeatedly moving said detectors relative to said observed image.

Claim 28

A method according to claim 25, wherein said dithering step further includes moving a gimbal mounting said detectors, thereby changing the angular position of said detectors.

Claim 29

A method according to claim 24, wherein said dithering step includes repeatedly changing the optical path of said detectors with a steerable mirror.

Claim 30

A method according to claim 24, wherein said dithering cycle includes a staring interval during which said observed image is stationary relative to said detectors.

Claim 31

A method according to claim 24, wherein said observed image is continuously dithered throughout said dithering cycle.

Claim 32

A method according to claim 24, wherein four adjacent detectors are traversed once during said dithering cycle.

Claim 33

A method according to claim 24, wherein said step of forming image signals includes the step of multiplying said image signals by a gain correction factor.

Claim 34

A method according to claim 24, wherein said filtering step further includes independently filtering each detector with a separate temporal high-pass filter.

Claim 35

A method according to claim 24, further includes the step of: periodically sampling each of said image signals at a period corresponding to an integer multiple of said dithering cycle, such that an output signal representing the observed image is formed.

Claim 36

A method according to claim 24, wherein said generating step further includes the step of convolving said filtered image signals with an inverse kernel of said temporal high-pass filter.

Claim 37

A method according to claim 34, further including the step of deriving said inverse kernel based on the line of sight motion of said detectors.

Claim 38

A method according to claim 24, further comprising the step of iteratively correcting the fixed pattern noise in said observed image as a function of said image signal, said reconstructed image, and an old noise signal representative of a previously determined fixed pattern noise in a previously observed image.

Claim 39

A method according to claim 36, wherein said correcting step comprises the steps of: generating an offset update signal corresponding to the difference between said reconstructed image and said image signal, forming a noise correction signal corresponding to sum of said generated offset update signal and said old noise signal, and offsetting said image signal by said noise correction signal.

Claim 40

A method according to claim 37, wherein said correcting step further comprises the step of modifying said offset update signal by a gain value.

Claim 41

A method according to claim 38, wherein said modifying step includes varying said gain value over time.

Claim 42

A method according to claim 37, further comprising the step of reducing errors in said noise correction signal by selectively processing said noise correction signal.

Claim 43

A method according to claim 40, wherein said error reducing step includes comparing the absolute value of successive noise correction signals and selectively transmitting a noise correction signal having a smaller absolute value.

Claim 44

A method according to claim 24 further comprising the step of averaging signals over successive frames to reduce temporal noise.

Claim 45

Apparatus according to claim 1 further comprising means for deriving said dithering cycle based on a difference between a first observed image observed by said array of detectors and a second observed image observed by said array of detectors at a time subsequent to said first observed image.

Claim 46

Apparatus according to claim 43 wherein said means for deriving said dithering cycle comprises means for shifting said second observed image relative to said first observed image, thereby creating a shifted second observed image; and means for computing a correlation between said first observed image and said shifted second observed image.

Claim 47

Apparatus for reducing fixed pattern noise in an image observed by an array of detectors, said apparatus comprising a plurality of image-responsive detectors, each said detector producing an image signal responsive to the portion of an observed image incident on said detector, thereby creating a set of image signals, dithering means for scanning said observed image across said detectors in a dithering cycle thereby forming a first observed image observed by said array of detectors and a second observed image observed by said array of detectors at a time subsequent to said first observed image, means for deriving said dithering cycle, said means including means for shifting said second observed image relative to said first observed image, thereby creating a shifted second observed image, and means for computing a correlation between said first observed image and said shifted second observed image, temporal high-pass filtering means having a non-zero DC response for generating a set of filtered image signals each of which corresponds to one of said image signals, each of said filtered image signals having an attenuated low frequency component, and image restoration means for generating a reconstructed image based on said set of filtered image signals, wherein said reconstructed image corresponds to said observed image having reduced fixed pattern noise.

Claim 48

Apparatus for reducing fixed pattern noise in an image observed by an array of detectors, said apparatus comprising a plurality of image-responsive detectors, each said detector producing an image signal responsive to the portion of an observed image incident on said detector, thereby creating a set of image signals, dithering means for scanning said observed image across said detectors in a dithering cycle thereby forming a first observed image observed by said array of detectors and a second observed image observed by said array of detectors at a time subsequent to said first observed image, means for deriving said dithering cycle, said means including means for shifting said second observed image relative to said first observed image, thereby creating a shifted second observed image, and means for computing a correlation between said first observed image and said shifted second observed image, temporal high-pass filtering means having a non-zero DC response for generating a set of filtered image signals each of which corresponds to one of said image signals, each of said filtered image signals having an attenuated low frequency component, and image restoration means for generating a reconstructed image based on said set of filtered image signals, wherein said reconstructed image corresponds to said observed image having reduced fixed pattern noise.