Dynamic Integration Time Adjustment for IR Camera Offset Map Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing IR camera technologies face challenges in maintaining optimal image quality across varying scenes due to fixed integration times, which lead to deviations in image quality and the generation of spatial fixed-pattern noise, especially in low-contrast scenes, and require frequent updates of offset maps through NUC processing.
Innovation Solution
A procedure that dynamically adapts integration time based on scene content by recording reference images at different integration times, updating the offset map, and using linear interpolation to create a compensation map, allowing continuous automatic adjustment without needing further NUC processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fixed integration times are used, then the offset map can be maintained stable, but image quality deviations occur and spatial fixed-pattern noise is generated
Solution Approach 1:
The patent implements dynamic integration time adjustment based on scene temperature characteristics. The system automatically selects integration times from multiple predefined positions (e.g., 5 different integration times) according to the actual scene conditions, transitioning from fixed to dynamic integration time management. This resolves the contradiction by allowing the system to adapt integration time to scene requirements while maintaining offset map stability through selective updates only when necessary.
Solution Approach 2:
The patent changes the integration time parameter dynamically based on scene temperature and content characteristics. By establishing multiple integration time positions and selecting appropriate ones according to actual scenes, the system optimizes image quality without requiring frequent offset map updates. The offset map is updated selectively based on temperature thresholds and scene change detection, rather than being tied to fixed integration time positions.
2Manufacturing precision
If multiple fixed integration times are introduced, then image quality is improved, but device complexity increases
Solution Approach 1:
Instead of maintaining separate offset maps for each fixed integration time position, the patent implements a dynamic selection mechanism that chooses from 5 predefined integration time positions based on scene characteristics. This reduces complexity by eliminating the need for multiple static map sets while maintaining image quality through adaptive integration time selection and selective NUC triggering.
Solution Approach 2:
The patent changes integration time dynamically based on scene temperature and content rather than using multiple fixed positions. By establishing a relationship between integration time and scene characteristics, the system selects appropriate integration times from a limited set of predefined positions, reducing the complexity of managing multiple fixed maps while maintaining optimal image quality.
3Manufacturing precision
If integration time is changed dynamically, then optimal image quality is achieved for different scenes, but low-frequency unevenness and spatial fixed-pattern noise increase
Solution Approach 1:
The patent implements feedback mechanisms to detect scene changes and temperature variations, triggering NUC (Non-Uniformity Correction) only when necessary. The system monitors scene characteristics and integrates this feedback with integration time selection, updating offset maps selectively rather than continuously. This feedback-based approach maintains image quality while minimizing spatial fixed-pattern noise by updating maps only when scene conditions warrant it.
Solution Approach 2:
The patent performs preliminary establishment of offset maps at multiple temperature points during production, creating a library of calibration data. During operation, the system selects from these pre-established maps based on current scene temperature, avoiding the need for continuous NUC processing. This preliminary action reduces spatial fixed-pattern noise by using stable, pre-calibrated maps while maintaining dynamic adaptation to scene conditions.
4Measurement precision
If NUC processing is performed frequently, then offset map accuracy is maintained, but loss of time and processing overhead increase
Solution Approach 1:
The patent changes the triggering condition for NUC processing from being tied to fixed integration time positions to being based on scene temperature thresholds and detected scene changes. The system performs NUC selectively when temperature variations exceed predefined thresholds or when significant scene changes are detected, rather than updating maps at every integration time change. This reduces processing time and overhead while maintaining offset map accuracy through intelligent, condition-based updates.
Solution Approach 2:
The patent implements feedback-based NUC triggering that monitors scene characteristics and temperature variations. The system determines whether NUC processing is necessary based on actual scene conditions rather than following a fixed schedule. This feedback mechanism maintains offset map accuracy by updating maps only when scene changes warrant it, significantly reducing unnecessary processing time and computational overhead.
Data Source
AI summary
The invention relates to a procedure for mapping when capturing video streams by means of a camera, such as an IR camera, as well as to a computer program and a computer program product. According to the procedure the following items are included: a) at least two reference images are recorded in production against a black-body radiator at the same temperature for two or more separate integration times (block 3), b) in conjunction with the updating of an offset map, details of the integration time for which the offset map has been updated are stored (block 6), c) during operation, the actual integration time is compared with the integration time for the most recent updating of the offset map (block 8), d) the recorded reference image which lies closest to the actual integration time is selected as a reference (block 7), and e) a compensation per pixel for the change in the integration time is calculated by linear interpolation between the selected reference image and the most recently updated offset map, resulting in a compensation map (block 8). The integration time can be adapted dynamically by the procedure with access to adequate maps for correction of the offset of constituent pixels.