Ecogrid: A serverless AWS solution for vegetation encroachment severity evaluation using contour detection and area ratio algorithm with open-source satellite imagery

Abstract

Vegetation encroachment is an issue faced by power grids due to trees and plant growth, which is harming grid components. This study proposes a system that takedown high-resolution satellite imagery from Google Earth Engine (GEE) based on target corridors established with Open Street Map (OSM) data. Effective imagery analysis results in a preprocessing step comprising cropping, color conversion, and additional inversion, which also analyzes the images by using contour detection and area calculations to quantify vegetation encroachment and categorize severity zones. With ReactJS and Leaflet, we build an interactive web dashboard for stakeholders to visualize these zones classified by the robust algorithm in real time to see the severity of vegetation. Data are stored and retrieved using a database and Flask Application Programming (API) in the backend. The research concluded that Serverless Image Processing Architecture with Amazon Web Services (AWS) Lambda for Cost Efficiency and Future Data Scalability is achieved. This will equip stakeholders with the information they need for better decisions when proactively managing vegetation. This monitoring system helps ensure the safe and reliable operation of the power grid.