Remote gas leak detection using the M3 multibeam sonar
Reliable detection of gas leaks in the marine environment is important from an environmental as well as an economic point of view. CCS (carbon capture and storage) has been proposed as a technology for significantly reducing greenhouse gas emissions, by capturing CO2 from industrial sources and permanently storing it in geological reservoirs. Reliable and preferably automatic monitoring of the marine environment above such a storage site is needed in order to verify that the injected CO2 is safely stored and that there are no signs of leakage.
Figur 1: Sonar frame/template with serval echo sounders and sonars mounted. The M3 is the scpherical black sonar to the left. Data from the other sonars and echo sounders may also be made available for this project.
The M3 sonar has been demonstrated as a promising technology for detailed monitoring of a potential risk area.
It would be interesting to further evaluate the potential that in particular the M3 sonar has, for automatic leak detection.
There are several possible approaches to this, including coherent and noncoherent change detection based on sonar imagery. This study would be based on raw M3 data acquired during a controlled CO2 release experiment in the North Sea this year. Figure 1 show the M3 sonar (black spherical sonar to the left) mounted on a seabed frame.
Figur 2: Sonar image A: standard processing has been used to image the seabed. The simulated CO2 bubble leak is visible at 65 m range, and another template (Template A) is visible at ~70 m range. A school of fish is also visible in this image.Figur 3: Sonar image B: A simple averaging filter used to enhance moving objects such as gas bubbles and fish, while suppressing stationary objects. Can we improve and automize this for more reliable gas leak detection?
Figures 2 and 3 show the M3 sonar image. Figure 4 shows the simulated CO2 leak, documented using a subsea video camera.
Figur 4: The simulated CO2 leak was documented using a subsea camera.
Necessary experience:
IN5450 - Array signal processing
INF4010 - Digital signal processing II
Solid Python / Matlab scientific programming and data analysis skills
