Thermal Anomaly Detection Using Statistical Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Interpreting thermal images from thermal imaging cameras requires specialized knowledge, as temperature readings alone are insufficient to determine normal or abnormal equipment operation, and fluctuations in thermal characteristics can lead to incorrect assumptions about equipment condition.
Innovation Solution
A system and method for detecting thermal anomalies by comparing acquired thermal image data to a statistical thermal profile, which includes mean, median, standard deviation, and percentile statistics, allowing for the identification of anomalous behavior in specific regions of the image, and using auxiliary test devices to account for parameters like current draw and ambient temperature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If temperature readings from thermal imagery are used to determine equipment condition, then thermal inspection can be performed quickly without invasive measurements or system downtime, but the interpretation requires substantial experience and specialized knowledge to distinguish normal from abnormal behavior
Solution Approach 1:
The patent introduces an intermediary system that includes a library of thermal profiles representing normal and abnormal equipment conditions. This intermediary acts as a mediator between the raw thermal imagery and the inspector, automatically comparing captured thermal data against stored profiles to identify anomalies. This eliminates the need for inspectors to have extensive specialized knowledge while maintaining high inspection speed.
Solution Approach 2:
The system enables self-service by allowing the thermal inspection system to automatically perform the interpretation function that previously required human expertise. The automated comparison of thermal profiles against the library and the generation of anomaly reports allow the system to serve itself, reducing dependency on highly trained inspectors while maintaining productivity.
2Measurement precision
If temperature readings are used without additional contextual information, then the measurement process is simple and quick, but the readings may not be sufficient to determine normal operation versus abnormal operation under varying load and ambient conditions
Solution Approach 1:
The patent applies preliminary action by pre-establishing a library of thermal profiles that capture the normal thermal characteristics of equipment under various operating conditions (different loads, ambient temperatures, etc.). These profiles are created and stored in advance, so when actual inspections occur, the system can immediately compare current readings against the appropriate pre-established profiles, ensuring accurate interpretation without needing to collect additional contextual information during the inspection itself.
3Measurement precision
If thermal characteristics of certain portions of equipment are monitored, then specific anomaly detection is possible, but these portions may have greater fluctuations during use than other portions, leading to incorrect assumptions about abnormal operation
Solution Approach 1:
The patent applies local quality by creating region-specific thermal profiles for different portions of equipment. Each region is characterized separately in the thermal profile library, capturing its normal fluctuation patterns under various operating conditions. During inspection, the system compares the thermal reading from each specific region against its corresponding regional profile, allowing the system to account for the fact that different regions have different normal fluctuation characteristics. This prevents false anomalies by comparing each region against its own baseline rather than a generic overall baseline.
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
Enables accurate detection of thermal anomalies, reducing the risk of misinterpretation by providing a graphical representation of anomalous regions and ensuring that temperature changes are evaluated within the context of typical operational conditions.
Implementation Method 1
A thermal imaging device can be used to acquire thermal image data of a target scene
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods can be used to detect thermal anomalies in a target scene of an infrared image. Acquired thermal image data can be compared to a statistical thermal profile to detect thermal anomalies in the image data. Anomaly data based on the comparison can be used to generate an image representing locations and/or severity of detected anomalies. Systems can be used to acquire thermal image data for generating and/or updating statistical thermal profiles for use in anomaly detection processes. Auxiliary measurement devices can provide measurement data representative of one or more parameters of the target scene. The measurement data can be used to select from a plurality of possible statistical thermal profiles associated with the target scene to best match the current parameters of the scene.