How AI-Powered Thermal Imaging Improves Leak Detection, Alarm Verification and Response

Print 🖨

The integration of artificial intelligence with fixed, continuous vision technology has transformed camera-based leak detection by enabling timely action through early detection and remote alarm verification. It targets certain coverage gaps challenging to address with standard industry practices, given the limitations of their fundamental design and objective.

Traditional vision methods, such boots on the ground, handheld cameras, aircraft/vehicle mounted, 24/7 manned surveillance cameras, have certain limitations:

Implemented to support non-visual methods (e.g., pressure/balance measurements and acoustic sensing) for alarm verification and maintenance inspections
Intermittent in nature and/or expensive to deploy more frequently or continuously
Partial visibility leaves visual monitoring gaps
Methods requiring manual operation at the site can expose personnel to hazards

Oil and gas facilities that are remote, irregularly staffed, and/or located near sensitive environments have a more comprehensive visual monitoring requirement. These applications necessitate vision technologies to be continuously operational, accurate and fast. Deficiencies in leak detection coverage can counter preventive or mitigation efforts, leading to greater impacts, such as environmental damage, financial losses and industry scrutiny.

Thermal Imaging with AI for Leak Detection

Thermal imagers produce greyscale images by interpreting the heat signature of objects. As a result, most product releases invisible to a color camera or the naked eye can be visualized with thermal cameras. These include small sprays and pooling, leaks that blend with the background and wet gasses like butane and propane. Advanced thermal imagers have higher resolutions and can capture radiometric data and scene changes at the pixel level.

How Vision AI Works:

Image processing AI software may include any combination of video analytics, general and application-specific detection algorithms, environmental filters, background learning, and deep learning object classification.
Edge-based AI analyzes live thermal video footage in real time, allowing greater processing capability with minimal network bandwidth, leading to higher accuracy.
Cloud processing requires continuous streaming and pre-transmission video compression, resulting in inferior data quality that may affect processing accuracy and high data costs.
Accuracy level and false alarm rates can be affected by the quality of the video, the data source, and the analytic software, but improve with updates.
Automates the repetitive and time-consuming process of monitoring and detection.
Customization capabilities and expertise of AI provider important success factors. Industry specific specializations and experience provide a competitive edge.

Dual sensor AI vision showing analytic tracking on the LWIR thermal camera of a simulated petroleum product release at 20-meter target distance and leak volume of .35L/s @ 80 psi.

Greater Automation, Visibility and Control

A multitude of tasks can be automated with AI-Powered Vision, leading to improved visibility, leak response, safety, productivity, and cost savings. With camera-based leak detection, energy operators gain the following capabilities and advantages:

Non-stop operation and condition monitoring of one or hundreds of distributed assets (limited by the capacity of the video management system)
Automatic leak detection and alerting, saving hundreds of manual monitoring time, effort and wages
Improved detection by having the ability to detect leaks up to 45 times smaller than CPM’s lower detection limits (equivalent to 0.1 percent nominal rate)
Remote access to visual data (live view, alarm package and recorded videos) to enable faster and safe verification of alarms, which minimizes confirmatory site visits, windshield time and personnel safety risks, and helps with stopping leaks early
Multi-region monitoring allows unique detection settings for different areas and assets within one camera view
Dual sensor view matches thermal and color video perspectives with the color camera used for verification
Dual sensor processing which applies analytics to both thermal and color sensors
Remote monitoring reduces expensive, lengthy and potentially risky periodic in-person inspections
Compatibility with other sensors and reporting interfaces to facilitate communication with operator control room, supervisory control and data acquisition (SCADA) system, and third-party monitoring services
On-going updates in software and hardware advancements to address new and evolving industry needs

IntelliView has been serving the oil and gas industry since 2011as a leading, industry recognized innovator and provider of AI-vision leak detection technology. Its DCAM system features thermal/color cameras and onboard AI specialized to detect oil, liquid and wet hydrocarbon leaks, and it comes with enclosure options (industrial grade housing or certified explosion-proof). The technology has been implemented at various oil and gas facilities, including pump and pigging stations, metering stations, terminals, refineries, well pads, and offshore platforms.

To know more, visit www.intelliviewtech.com or contact IntelliView.