Infrared Gas Imaging with Location-Specific Emission Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to accurately quantify fugitive emissions due to environmental variability and 'burn-in' effects, leading to improper detection and quantification of gas leaks, especially when the field of view of infrared imaging devices changes.
Innovation Solution
Implementing a hyperspectral camera with an infrared imaging device that generates IR image data and a computing device to access location-specific detection models, iteratively updated to account for changing fields of view, reducing computational burden and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single common model is used for gas detection at all locations, then device complexity is reduced, but measurement precision deteriorates due to environmental variability and burn-in effects
Solution Approach 1:
The patent divides the detection system into multiple location-specific models, each trained for a specific field of view location. This segmentation allows each model to specialize in detecting gas emissions at its particular location, accounting for local environmental conditions and burn-in effects, thereby improving measurement precision without requiring a single overly complex universal model.
Solution Approach 2:
The patent implements location-specific detection models where each model is tailored to the characteristics of its specific location. Each model learns the local background conditions, temperature profiles, and burn-in patterns specific to its field of view location, enabling more accurate gas detection at each location rather than using a generic model for all locations.
2Measurement precision
If location-specific detection models are implemented, then measurement precision improves, but device complexity increases due to multiple models and iterative updates
Solution Approach 1:
The patent performs preliminary training of location-specific detection models during idle periods or before actual gas detection operations. By pre-training models on historical data and environmental conditions for each location, the system prepares the models in advance, reducing the computational burden during real-time detection and making the deployment of multiple models more manageable.
Solution Approach 2:
The patent implements periodic updates of location-specific detection models during idle periods between gas detection operations. Instead of continuously updating models during active detection, the system periodically retrains and refines models using accumulated data, balancing measurement precision improvement with reduced computational complexity during critical detection phases.
3Productivity
If continuous imaging and model updates are performed, then productivity improves through automated detection, but use of energy increases due to computational burden
Solution Approach 1:
The patent performs model updates periodically during idle periods rather than continuously during detection operations. This periodic updating approach maintains automated detection productivity while significantly reducing energy consumption by avoiding constant computational processing during active gas detection phases.
Solution Approach 2:
The patent maintains continuous automated gas detection capability while performing model updates during idle periods. The system ensures continuous monitoring and detection readiness without requiring continuous high-energy computational processing, as models are refreshed periodically when detection demands are lower, thus maintaining productivity with reduced energy overhead.
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
Provides continuous, automated, and accurate detection and quantification of fugitive emissions by leveraging location-specific models, minimizing computational load and accounting for environmental variations.
Implementation Method 1
an infrared (IR) imaging device configured to generate second IR image data of a first field of view of the IR imaging device at a second time
Implementation Method 2
generate first spectral absorption data based upon the second IR image data and the first detection model and detect a fugitive emission within the first field of view based upon the first spectral absorption data
Data Source
AI summary
Systems, methods, and computer program products for multi-model emission determinations are provided. An example imaging system includes an infrared (IR) imaging device configured to generate second IR image data of a first field of view of the IR imaging device at a second time and a computing device operably connected with the IR imaging device. The computing device receives the second IR image data of the first field of view from the IR imaging device and accesses a first detection model associated with the first field of view of the IR imaging device. The first detection model is generated based upon first IR image data of the first field of view of the IR imaging device generated at a first time. The computing device further generates first spectral absorption data based upon the second IR image data and the first detection model for detecting a fugitive emission.