Infrared Detector Array Non-Uniformity Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Infrared detector arrays suffer from non-uniformity noise due to variations in dark current, offset, responsivity, and other factors, which affect image quality and require frequent recalibration, especially over a wide temperature range.
Innovation Solution
A method involving a focal plane array of photodetector elements that captures two images with different integration periods and computes their difference to generate a corrected image, effectively canceling out non-uniformity factors independent of integration time, with stable gain coefficients that require only one-time calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If two-point correction techniques employing reference images at two different temperatures are used, then non-uniformity correction is valid over a wider temperature range, but the system requires frequent recalibration and complex mechanical systems
Solution Approach 1:
The patent changes the integration period parameter to create different images with distinct temporal characteristics. By subtracting images with different integration periods, the method isolates and eliminates non-uniformity components that are independent of integration time, thereby achieving temperature-range validity without frequent recalibration
Solution Approach 2:
The patent uses copies of the same scene captured at different integration periods as reference images. Instead of requiring physical reference objects at different temperatures, the system creates virtual reference images by capturing the same scene multiple times with varying integration periods, eliminating the need for complex mechanical calibration systems
2Measurement precision
If offset coefficients are used for non-uniformity correction, then image non-uniformity is compensated, but the coefficients drift and require frequent recalibration
Solution Approach 1:
The patent dynamically adjusts the integration period between captured images to create a temporal difference that isolates non-uniformity components. This dynamic approach allows the system to continuously compensate for drift without requiring static, unstable offset coefficients to be manually recalibrated
Solution Approach 2:
The system uses its own captured images at different integration periods as the correction reference, eliminating the need for external calibration sources. The images themselves serve as the correction data, allowing the system to self-calibrate and maintain coefficient stability over time
3Ease of manufacture
If image blurring is used in calibration, then reference images can be obtained, but scene details are lost in the corrected image
Solution Approach 1:
The patent segments the image correction process into two independent steps: first capturing images at different integration periods to isolate non-uniformity, then subtracting these images to eliminate non-uniformity while preserving all scene details. This segmentation avoids the need for blurring during calibration while achieving both reference image acquisition and detail preservation
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces overall non-uniformity noise in images while preserving scene details, eliminating the need for frequent recalibration and complex mechanical systems, and maintains image quality across varying temperatures.
Implementation Method 1
a focal plane array of photodetector elements for detecting electromagnetic radiation from a scene
Data Source
AI summary
Methods and apparatus for effecting a non-uniformity correction of images of a scene obtained with an array of detector elements are disclosed. A first image of the scene having a first integration period is acquired using the array of detector elements. A second image of the scene having a different integration period is acquired, and a corrected image of the scene is generated by computing a difference of the images. In some embodiments, the first and second images are images of substantially identical scenes. According to some embodiments, the images are infrared images. Optionally, the corrected image is subjected to further correction using pixel dependent correction coefficients, such as gain coefficients. Exemplary image detection elements include but are not limited to InSb detector elements and ternary detector elements, such as InAlSb, MCT (Mercury Cadmium Telluride), and QWIP technology (Quantum Well Infrared Photodiodes). In some embodiments, the detector elements are cooled to a temperature substantially equal to an atmospheric boiling point of liquid nitrogen. Alternatively, the detector elements are cooled to a temperature below an atmospheric boiling point of liquid nitrogen, or any other operating temperature.