Infrastructure Monitoring using Satellite Imagery and Computer Vision

Abstract

Infrastructures like roadways and power lines are essential to our society’s economy and quality of life, yet they are vulnerable to the environment and severe weather. Trees falling on power lines or obstructing roadways are typical examples of how such vulnerabilities can affect society.

These situations highlight the need for effective and efficient monitoring of infrastructure. Traditional monitoring methods, like ground patrols and visual inspections, are resource-intensive, particularly in vegetation management. These methods are costly, time-consuming, and pose safety risks and environmental concerns. Moreover, the difficulty of monitoring infrastructure is exacerbated by the extensive scale of these networks, especially power lines and roadways, which often span hundreds or even thousands of kilometers. Consequently, the infrequency of inspections often leads to reliance on outdated data, which can hinder the process of making timely and effective decisions.

High-resolution satellite imagery offers a more efficient, cost-effective, and environmentally friendly alternative to traditional methods, providing frequent and comprehensive coverage. This approach enables proactive infrastructure monitoring, improving situational awareness and facilitating prompt action.

This research is centered on developing methodologies and algorithms within the fields of computer vision and earth observation using the most recent remote sensing technologies for infrastructure monitoring. Such methodologies are designed for analyzing satellite imagery to characterize vegetation, detect changes, assess risks, and offer current updates on the condition of infrastructure systems. It covers two main areas: monitoring power line infrastructure to characterize vegetation encroachment and ensure operational integrity and monitoring roadways in natural disaster management. By utilizing satellite imagery, the research aims to offer more timely, cost-effective, and comprehensive monitoring, significantly improving traditional methods and enhancing infrastructure resilience.