Probabilistic Method to Fuse Artificial Intelligence-Generated Underground Utility Mapping.

Area of Science:

• Geospatial Engineering
• Geophysics
• Data Science

Background:

• As-built utility plans often contain inaccuracies, leading to accidental strikes during excavation.
• Existing utility investigation methods struggle with data interpretation complexities and high costs, especially for large projects.
• Data fusion offers improved accuracy but faces limitations in practical application.

Purpose of the Study:

• To develop a novel framework for accurate and cost-effective large-scale utility mapping.
• To address challenges in data interpretation and reduce the cost of utility investigations.
• To generate probabilistic inferences for identifying high-uncertainty regions requiring further study.

Main Methods:

• Automatic initial map creation using utility inference rules on identified appurtenances.
• Data fusion of the initial map with as-built data or historical satellite imagery.
• Uncertainty evaluation using confidence value estimators to produce a probabilistic utility map.

Main Results:

• A novel framework for rapid, low-cost utility infrastructure mapping.
• Probabilistic utility maps revealing confidence levels for buried utility locations.
• Identification of high-uncertainty regions as targets for focused investigations.

Conclusions:

• The proposed framework enhances the accuracy of underground utility mapping by providing probabilistic inferences.
• It significantly reduces costs and improves efficiency by limiting detailed investigations to critical areas.
• The dynamic nature of the framework allows for automatic updates, ensuring long-term data relevance and minimizing obsolescence.