Scene-Based Non-Uniformity Correction for Infrared Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Infrared imaging systems, particularly those using microbolometer sensors, face challenges with fixed pattern noise (FPN) that conventional calibration methods struggle to fully address, leading to residual noise even after non-uniformity correction, due to temperature differences and non-linearity issues between the shutter and scene temperatures.
Innovation Solution
The implementation of a scene-based non-uniformity correction method that involves generating a template frame, determining frame-to-frame motion, comparing pixel data to identify irradiance differences, propagating pixel offset information for correction, and updating correction terms to reduce FPN in infrared images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional calibration algorithms with internal shutter are used to minimize non-linearity and temperature effects, then some non-uniformity correction is achieved, but residual fixed pattern noise remains due to temperature differences and non-linearity between shutter and scene temperatures
Solution Approach 1:
The patent changes the fundamental parameter being measured from shutter temperature to scene temperature. By detecting irradiance from the scene and using frame-to-frame motion to propagate corrections, the system adapts to actual scene conditions rather than relying on shutter temperature assumptions, thereby reducing residual FPN caused by temperature mismatches
Solution Approach 2:
The patent replaces the mechanical shutter-based calibration system with a scene-based optical measurement system. Instead of using a mechanical shutter to provide uniform reference, the system uses scene irradiance detection combined with motion compensation to achieve correction, eliminating the fundamental limitation of shutter temperature vs. scene temperature differences
2Ease of manufacture
If one point or two point correction is applied at shutter temperature, then offset and gain corrections are achieved, but corrections are not appropriate for scenes with different temperatures due to detector non-linearity
Solution Approach 1:
The patent transforms the static shutter temperature correction into a dynamic scene-adaptive correction system. By continuously detecting scene irradiance and using frame-to-frame motion to update correction terms, the system dynamically adapts to varying scene temperatures and conditions, overcoming the fixed temperature limitation of conventional methods
Solution Approach 2:
The system uses the scene itself as the reference for correction. By detecting irradiance from the scene and using the scene's own motion patterns, the system performs self-calibration without requiring external shutter mechanisms, making the correction inherently appropriate for the actual scene being imaged
3Manufacturing precision
If internal shutter calibration is used to acquire image against uniform target, then non-uniformity correction is performed, but residual noise remains because corrections only address sources between shutter and detector
Solution Approach 1:
The patent extracts the calibration reference from the mechanical shutter and relocates it to the scene itself. By using scene irradiance as the reference and propagating corrections through detected frame-to-frame motion, the system addresses non-uniformity sources throughout the entire optical path including the lens and scene interface, not just between shutter and detector
Data Source
AI summary
Systems and methods provide scene-based non-uniformity correction for infrared images, in accordance with one or more embodiments. For example in one embodiment, a method of processing infrared images of a scene captured by an infrared image sensor comprising a plurality of sensor elements includes receiving a first frame comprising a first plurality of pixel data of a first infrared image; receiving a second frame comprising a second plurality of pixel data of a second infrared image; determining frame-to-frame motion between the first frame and the second frame, wherein the frame-to-frame motion identifies portions of the first and second pixel data corresponding to identical scene coordinates captured by different sensor elements for the first and second frames; determining irradiance differences between the first and second portions of pixel data; and determining pixel offset information for scene based non-uniformity correction terms based on the irradiance differences and the frame-to-frame motion.