Artificial Intelligence Models for Digitized Operations and Maintenance of Large Infrastructure Systems
Description
Large-scale civil infrastructure systems are critical for the functioning and development of any society. However, these systems are often vulnerable to degradation and the effects of aging, necessitating consistent monitoring and maintenance. Current methods for infrastructure maintenance primarily rely on human intervention and need the implementation of advanced sensing and computing technologies in field operations and maintenance (O&M) tasks. This research aimed to address these gaps and provide novel contributions. Specifically, the objectives of this study were to leverage artificial intelligence models to enhance point cloud noise processing, to automate tree species detection using Mask R-CNN, and to integrate imagery data and LiDAR datasets for real-time terrain analysis. First, the study proposed leverages neural networks to eliminate unwanted noise from point cloud datasets, enhancing the accuracy and reliability of infrastructure data. Secondly, the research integrated Mask R-CNN into automated tree species detection. This component offers an efficient solution to identify and classify vegetation surrounding infrastructure, enabling infrastructure managers to devise proactive vegetation management strategies, thereby reducing risks associated with tree-related incidents. Lastly, the study fused image and LiDAR datasets to support real-time terrain analysis. This integrated approach provides a comprehensive understanding of terrain characteristics, allowing infrastructure managers to assess slope, elevation, and other relevant factors, facilitating proactive maintenance interventions and mitigating risks associated with erosion. These contributions collectively underscore the potential of artificial intelligence models in advancing the operations and maintenance practices of large civil infrastructure systems. By leveraging these models, infrastructure managers can optimize decision-making processes, streamline maintenance efforts, and enhance critical infrastructure networks' overall resilience and sustainability.
Date Created
The date the item was original created (prior to any relationship with the ASU Digital Repositories.)
2023
Agent
- Author (aut): Paladugu, Bala Sai Krishna
- Thesis advisor (ths): Grau, David
- Committee member: Ernzen, James
- Committee member: Standage, Richard
- Publisher (pbl): Arizona State University