Infrared imaging technique detects vapor leaks

Engineers atProvidence Engineering (Baton Rouge, LA, USA) have been awarded a US patent (8,009,932 B2) for developing an infrared imaging and image-analysis technique that has been deployed in a system to detect hydrocarbon leaks from process equipment.

Currently, technicians manually examine individual process components for leaks using handheld instruments. While the method is widely used to comply with US Environmental Protection Agency (EPA) Leak Detection and Repair (LDAR) regulations, it is labor intensive due to the large number of components at a refinery or chemical plant that must be checked on a quarterly basis.

As a result, the industry has been working with the EPA to develop a new method called "Smart LDAR." Smart LDAR systems use an infrared (IR) camera that can image the hydrocarbon vapors that would not otherwise be visible. As such, they allow a technician to check a much larger number of components in a day.

The Providence engineers have automated that process by developing image-processing software that can automatically isolate, recognize, and/or quantitatively analyze vapor plume images captured by the IR camera, in a system that they call third-generation LDAR, or LDAR3.

The newly patented technique, which can automatically align image frames in recorded video clips, is one of the enabling technologies behind the LDAR3 system. Specifically, it uses spatially based fast Fourier transforms (FFTs) to determine the amount of shift, rotation, and scaling required to align image frames, and uses a digital filtering technique to improve the alignment.

The result is that it is now possible to analyze individual frames in the video images that were previously difficult to process further due to the fact that they may have shifted or rotated as a result of shaking or vibration of the IR camera.

-- Posted byVision Systems Design