Thermal Imaging Fixed Pattern Noise Correction Without Shutter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional thermal imaging techniques face challenges in effectively correcting fixed pattern noise (FPN) without requiring calibration shutters, which adds complexity and is not effective across a broad range of spatial noise frequencies.
Innovation Solution
A method and device that calculate pixel difference values between a temporal IR-image-average and captured IR images, evaluating these differences to update fixed pattern noise correction terms, allowing for FPN correction in thermal images without calibration shutters, even when no initial correction image is available, and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional non-uniformity correction (NUC) techniques are used to correct fixed pattern noise, then FPN correction can be achieved, but a calibration shutter is required causing video image freezing and adding complexity to imaging systems
Solution Approach 1:
The patent extracts the FPN correction function from the conventional NUC process that requires a calibration shutter. By separating the FPN correction from the shutter-based calibration mechanism, the system achieves FPN correction without requiring physical shutter components, thereby reducing device complexity while maintaining correction effectiveness
Solution Approach 2:
The patent replaces the mechanical calibration shutter system with a computational approach using temporal filtering and pixel difference analysis. Instead of using a physical shutter to capture reference images for calibration, the system uses software-based processing of continuous video frames to calculate and apply FPN correction terms, eliminating moving mechanical parts
2Reliability
If conventional NUC techniques are used for FPN correction, then correction can be performed, but the system requires calibration shutter closure which freezes video imaging
Solution Approach 1:
The patent enables continuous video imaging by performing FPN correction on each frame as it is captured, without requiring interruption for shutter closure. The temporal filtering process continuously processes incoming frames, calculating pixel difference values and updating correction terms in real-time, ensuring uninterrupted video production while maintaining FPN correction
Solution Approach 2:
The patent performs preliminary calculations of pixel difference values between the current frame and the temporal average before applying correction. By pre-calculating these difference values and using them to determine correction terms, the system prepares correction data in advance for each frame, enabling continuous processing without shutter interruptions
Data Source
AI summary
A method and a thermal imaging device for correcting fixed pattern noise of in a video sequence of thermal images captured using a thermal imaging system are provided. The method includes receiving a sequence of thermal images captured using a thermal imaging system. The method also includes calculating a thermal image average of a plurality of the received thermal images and calculating pixel difference values between a received thermal image and the thermal image average. The method evaluates the pixel difference values in relation to an updating condition and calculates fixed pattern noise correction terms for the pixel difference values for which the updating condition is fulfilled. The method also includes correcting pixel values of a subsequently received thermal image with the fixed pattern noise correction terms to generate a sequence of thermal images having corrected fixed pattern noise.