Advanced algorithms in Visual inspection of Pipeline
2018-10-30

The Sembcorp Industries Service Corridor is a 25 km long network of complex pipelines offering companies located in Jurong island’s Banyan, Tembusu and Angsana areas a comprehensive range of energy, water and on-site logistics. This network of pipelines transfers process steam, power, industrial wastewater treatment, industrial waters and service corridor services across these companies. The current manpower-intensive inspection of these myriad of pipelines usually takes a year to complete by manual human inspection with devices.

Looking for solutions that will alleviate the currently manpower-intensive inspection of each individual pipe in the Sembcorp Industries Service Corridor, the idea of using drones for autonomous visual inspection sprung.It is estimated that such visual inspection could possibly cut the time down to 2-3 weeks. Under the Sembcorp-EMA Energy Technology Partnership (SEETP) grant, A*STAR’s Institute for Infocomm Research (I2R) together with DNV-GL and Flare Dynamics Pte Ltd will thus be building up technologies needed in this Automated Pipeline Monitoring with Unmanned Aerial Imaging System.

A*STAR’s Institute for Infocomm Research (I2R) will be developing intelligent learning algorithms that will detect leakages automatically from the images and videos. It will also integrate this data with the UAV telemetry information to pinpoint the leakage location for human verification of the leaks.

The proposed solution consists of integration of the following key technologies:

  • Fully autonomous drone that is capable of 24/7 deployment.
  • Multiple Drone-mounted Sensors to variations along the pipelines and its environment due to leaks.
  • Developing image and video analysis algorithms for image processing to detect leakages automatically
  • Data integration between multi spectrum imaging and the UAV telemetry information (eg. GPS, Euler Angle etc) to pinpoint the leakage location and generate flight path accordingly.
  • Automatically flags potential leaks to the operator for further review

Gas Leakage detection using A*STAR I2R advanced video analytics algorithms


Read More below: 

Press Release: EMA in S$10 million Partnership Renewal with Sembcorp for R&D and Manpower Development