Issuu

International Research Journal of Engineering and Technology (IRJET)

e-ISSN: 2395-0056

Volume: 12 Issue: 12 | Dec 2025

p-ISSN: 2395-0072

www.irjet.net

Border Surveillance Using Satellite Imagery and Artificial Intelligence Prof. Shobha S. Biradar¹, Shreyas Gururaj Babaleshwar² ¹Professor, Master of Computer Applications, VTU CPGS, Kalaburagi, Karnataka, India ²Student, Master of Computer Applications, VTU CPGS, Kalaburagi, Karnataka, India ---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Border surveillance has become a critical

selected Area of Interest (AOI) and performs vegetation change mapping, movement analysis, object detection, footprint tracing, desert sandstorm suppression, and predictive risk modelling. The objective is to provide border security forces with accessible, real-time insights that support operational readiness, strategic decision- making, and preventive security interventions.

national security requirement due to increasing concerns regarding illegal migration, unauthorized crossings, smuggling, and strategic threats. This research proposes a satellite-based, artificial intelligence-driven surveillance system capable of continuously monitoring border regions using multi-temporal remote sensing data. The system integrates Google Earth Engine with intelligent modules for vegetation change analysis (NDVI), movement tracking, object detection (vehicles, buildings, and cargo containers), human footprint tracing, and predictive analytics for smuggling and migration risks. The proposed solution automates geospatial analysis through indices (NDVI, NDBI, NDWI, BSI), SAR backscatter, and temporal differencing. A Stream lit-based dashboard provides real-time analysis, alerts, and risk visualization for border forces. The system is lightweight, scalable, and low-cost, making it suitable for large border surveillance operations. Experimental evaluation shows significant potential for real-time border monitoring, early anomaly identification, and predictive security analytics.

1.1 Problem Identification Current border surveillance systems lack automated, largescale, terrain-independent reconnaissance capabilities. Manual methods are labor-intensive and fail to detect subtle or low-visibility activities such as gradual encroachment, illegal trails, temporary shelters, or night-time movements. The absence of automated anomaly detection exposes borders to security vulnerabilities. 1.2 Objectives of the Study

Key Words: Satellite Imagery, Google Earth Engine, NDVI, Border Surveillance, AI Analytics, Movement Detection, Predictive Risk Modeling, Object Detection

1. INTRODUCTION

Border surveillance plays an essential role in national security, especially for countries with extensive geopolitical boundaries. Traditional surveillance techniques depend heavily on human patrolling, CCTV networks, and local sensor systems. These approaches are limited in range, prone to human error, and often infeasible in remote terrains such as deserts, mountains, marshlands, and dense forests. Modern technological advancements have enabled the use of satellite imagery in conjunction with AI to monitor borders at scale, regardless of terrain or climatic conditions. Satellite missions like Sentinel-2 (optical) and Sentinel-1 (SAR) offer high-resolution, frequent, and cloud-penetrating imagery that can reveal surface anomalies, movement patterns, construction activities, and environmental changes. Artificial Intelligence further elevates the value of satellite data by providing automated detection, classification, and prediction capabilities.

1.3 Scope The project can be applied to: 1. 2. 3. 4. 5.

Impact Factor value: 8.315

International land borders Desert / coastal / mountainous terrains Defense intelligence operations Disaster response and illegal settlement monitoring Smuggling and infiltration detection

1.4 Methodology Overview Data is retrieved from Sentinel-1 and Sentinel-2 satellite collections. Various indices such as NDVI, NDWI, NDBI, and BSI are computed, along with RGB differencing for movement tracking. SAR backscatter identifies metallic or rigid structures. A predictive model computes risk levels

This research introduces a unified border surveillance system using Google Earth Engine (GEE) and AI-based analytical modules, accessible through a user-friendly Streamlit dashboard. The proposed system analyses a

To develop a satellite-driven AI surveillance dashboard for border monitoring. To implement change detection modules including NDVI, movement tracking, and dust storm masking. To detect vehicles, buildings, cargo containers, and footprints using Earth Engine imagery. To integrate predictive analytics for migration or smuggling-related risk assessment. To provide real-time alerting and intuitive geospatial visualization for authorities.

ISO 9001:2008 Certified Journal

Page 72