Infrared Image Deblurring via Thermal Blur Kernel Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing algorithms developed for visible light images often degrade in performance when applied to infrared images, leading to ineffective image deblurring due to domain dependence, which necessitates the development of algorithms suitable for the infrared image domain.
Innovation Solution
A method and apparatus for generating a temperature blur kernel using performance information from infrared sensors and gyro sensors, based on a heat balance equation, to correct image blurring in infrared images by obtaining output image equations and applying homography transform functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If general image processing algorithms developed for visible light images are applied to infrared images, then the algorithms can be used across different domains, but the performance degrades due to domain dependence
Solution Approach 1:
The patent changes the fundamental parameters of the blur kernel generation by using thermal time constant (τ) instead of exposure time, and by incorporating gyro sensor data (angular velocity ω) into the blur model. This transforms the blur kernel from a visible-light-based model to an infrared-specific model that accounts for thermal diffusion characteristics, thereby resolving the domain dependence issue while maintaining algorithm versatility.
Solution Approach 2:
The patent substitutes the mechanical/optical blur model (based on exposure time and aperture) with a thermal diffusion model (based on heat balance equation and thermal time constant). This replacement allows the deblurring algorithm to accurately model infrared image blur characteristics, which are governed by thermal processes rather than optical processes, thus improving reliability in the infrared domain.
2Reliability
If thermal time constant and angular velocity are incorporated into the blur kernel generation, then image deblurring performance improves, but the complexity of the processing system increases
Solution Approach 1:
The patent achieves multi-functionality by integrating multiple data sources (IR sensor, gyro sensor, and existing performance parameters) into a unified blur kernel generation framework. The same processing pipeline handles both thermal diffusion effects and motion blur effects, eliminating the need for separate processing modules and thus managing system complexity while improving deblurring performance.
Solution Approach 2:
The patent merges the thermal time constant parameter (from heat balance equation) with the angular velocity parameter (from gyro sensor) into a single integrated blur kernel calculation. This combination allows the system to account for both thermal diffusion and motion blur simultaneously in one processing step, reducing overall system complexity compared to handling these effects separately.
3Measurement precision
If the blur kernel is generated using heat balance equation and gyro sensor data, then the temperature blur kernel accurately reflects infrared image characteristics, but the computational requirements increase
Solution Approach 1:
The patent performs preliminary computation by pre-calculating the thermal time constant (τ) and incorporating it into the blur kernel generation formula. This preliminary action allows the system to use simple exponential decay functions during actual image processing, rather than performing complex thermal diffusion simulations in real-time, thus reducing computational power requirements while maintaining blur kernel accuracy.
Solution Approach 2:
The patent transforms the complex thermal diffusion problem into a simpler parameter-based solution by changing from a differential equation approach to an explicit parameter substitution approach. By using the thermal time constant as a direct parameter in the blur kernel formula (kt[x, y] = exp(-x²/(2σ²)) * exp(-y²/(2σ²)) where σ relates to τ), the system achieves accurate blur modeling with reduced computational complexity compared to full thermal diffusion simulation.
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
The proposed solution effectively generates a temperature blur kernel that improves image deblurring performance in infrared images by considering thermal time constants and angular velocities, resulting in higher peak signal-to-noise ratios and structural similarity compared to existing algorithms.
Implementation Method 1
obtaining an output image equation based on a heat balance equation
Implementation Method 2
obtaining a camera angle information by using a gyro sensor
Data Source
AI summary
The disclosure relates to a technique for generating a temperature blur kernel by using a heat balance equation. A method of generating a temperature blur kernel includes obtaining first performance information corresponding to an infrared (IR) sensor and second performance information corresponding to a gyro sensor, obtaining an output image equation based on a heat balance equation, the first performance information, and the second performance information, and generating a temperature blur kernel based on the output image equation, the first performance information, and the second performance information.