Variable-baseline disparity matching for 3D measurement and buried pipeline depth estimation

Area of Science:

• Geomatics Engineering
• Geotechnical Engineering
• Remote Sensing

Background:

• Existing underground pipeline surveys are costly, time-consuming, and limited in scope.
• Accurate 3D ground surface data is crucial for post-earthquake infrastructure assessment.
• Monitoring buried pipelines before and after seismic events is challenging.

Purpose of the Study:

• To develop a rapid 3D ground surface measurement technique using aerial imagery.
• To estimate the depth of shallow buried pipelines post-earthquake.
• To overcome limitations of traditional underground pipeline detection methods.

Main Methods:

• Utilized disparity matching of two aerial images to create 3D ground surface maps.
• Developed a small-scale observation system to simulate seismic events and measure pipeline responses.
• Applied Random Forest Regression for estimating buried pipeline depth from 3D ground data.

Main Results:

• Achieved ground displacement measurement errors under 2 mm and ground height errors under 1 cm.
• Demonstrated fast, wide-area 3D mapping capability in simulated disaster-affected regions.
• Obtained a pipeline depth prediction error of less than 5 mm using regression analysis.

Conclusions:

• The proposed aerial imaging method offers an efficient solution for 3D ground mapping in disaster areas.
• The system enables unprecedented data acquisition for pre- and post-earthquake pipeline monitoring.
• This approach significantly improves the accuracy and efficiency of assessing buried infrastructure after seismic events.