Reflection-Reduced Gas Imaging via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In gas facilities, the high-temperature flare stack emits a large amount of infrared rays, causing high-luminance reflections that interfere with gas leakage detection, making it difficult to observe changes in infrared rays due to the detection target gas and significantly lowering the gas detection rate.
Innovation Solution
A reflection-component-reduced image generating device that uses a machine-learned estimation model to reduce the image component of reflected light in gas distribution images, utilizing a combination of images with and without high-luminance light sources to generate a reflection-component-reduced image, thereby minimizing the influence of high-luminance light sources on gas visualization imaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If infrared moving image capture is used for gas leakage detection, then gas can be visualized and detection becomes easier, but high-luminance reflection components from flare stacks interfere with detection and lower the gas detection rate
Solution Approach 1:
The invention extracts and removes the high-luminance reflection component from the infrared image through image processing. By separating the reflection component (caused by flare stack illumination) from the gas signal, the system maintains the ease of gas visualization while eliminating the interference that reduced detection reliability
Solution Approach 2:
The invention introduces an intermediary processing step that models and subtracts the reflection component. This intermediary process acts as a mediator between the raw infrared image and the final gas detection result, removing the harmful reflection effects while preserving the useful gas visualization information
2Loss of energy
If flare stack is used for gas combustion, then surplus gas can be detoxified, but the high-temperature flame emits large amount of infrared rays that create high-luminance reflections and reduce detection accuracy
Solution Approach 1:
The invention converts the harmful effect of the flare stack's infrared emission into a manageable parameter. By modeling the reflection component based on the known flare stack illumination pattern and subtracting it from the image, the system eliminates the interference while maintaining the necessary gas combustion function for surplus gas disposal
3Illumination intensity
If high-luminance light source illuminates the target, then reflected light provides image information, but the changing amount of reflected light interferes with observing detection target gas and lowers detection rate
Solution Approach 1:
The invention extracts the time-varying reflection component caused by the high-luminance light source and removes it from the image signal. This allows the system to maintain adequate illumination for image formation while eliminating the interfering reflected light that obscured the detection target gas
Solution Approach 2:
The invention performs preliminary processing to predict and remove the reflection component before gas detection analysis. By anticipating and subtracting the interference pattern in advance, the system prevents the reflected light from masking the gas signal, thereby maintaining detection reliability
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
This approach improves the detection quality of gas leakage by reducing the interference from high-luminance light sources, allowing for more accurate visualization and detection of gas leaks.
Implementation Method 1
a risk of gas leakage is recognized due to aged deterioration of facilities and operational errors... an optical gas leakage detection method has been employed in which an infrared moving image is captured using infrared absorption characteristics of gas
Implementation Method 2
an optical gas leakage detection method has been employed in which an infrared moving image is captured using infrared absorption characteristics of gas to detect gas leakage
Implementation Method 3
equipment around the flare stack is illuminated by emitted infrared rays and is observed as a high-luminance reflection component
Data Source
AI summary
There are provided an inspection image input unit that receives a gas distribution image as an input, the gas distribution image having a visualized presence region of a gas in a space and including an image portion in which a target is irradiated with light, and a reflection-component-reduced image generating unit that generates a reflection-component-reduced image in which an image component of reflected light in the image portion of the gas distribution image received by the inspection image input unit is reduced using an estimation model machine-learned using, as teacher data, a combination of a first image including an image portion in which a target is irradiated with light and a second image including an image portion in which the target is not irradiated with light, the second image being equivalent to the first image for elements other than the image portion.