Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{195530,
        author = {Janapati Venkata Sriveda Manaswi and Nitturu Asha Jyothi and Janapati Venkata Sai Chandra Koustubh and Vartha Venu and Dasareddygari Jaswanth Reddy},
        title = {Smart Surveillance of Harmful Highway Vegetation Using Deep Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {723-734},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=195530},
        abstract = {The rapid spread of harmful and invasive roadside vegetation poses significant risks to highway infrastructure, road safety, and ecological balance. Traditional vegetation monitoring methods rely on manual inspection, remote sensing, or citizen-reported observations, which often suffer from limited coverage, high operational costs, and delayed response times. This study presents a Smart Surveillance System for Harmful Highway Vegetation using deep learning and edge computing to enable automated, geotagged detection from roadside video streams. The proposed system integrates a transformer-based deep learning model with a CPU-optimized edge inference pipeline, GPS-based geolocation tagging, and a web-based monitoring platform that separates administrative control from field-level access. By combining computer vision with spatial mapping, the system aims to provide scalable and efficient vegetation monitoring along transport corridors.
The methodology involves frame sampling from vehicle-mounted video capture, image tiling for high-resolution processing, deep learning–based binary classification, and threshold-based detection logging. Each detection is associated with geographic coordinates and visualized through an interactive web interface using geospatial mapping tools. Experimental evaluation conducted on Indian highway footage produced stable detection outputs with consistent confidence scores and reliable GPS tagging. The results demonstrate the feasibility of deploying a lightweight, edge-based vegetation surveillance framework capable of near real-time monitoring without GPU dependency. The proposed system contributes toward intelligent infrastructure management and supports future expansion into multi-class vegetation identification and large-scale deployment across transportation networks.},
        keywords = {Highway Vegetation Monitoring, Invasive Plant Detection, Deep Learning, Edge Computing, Roadside Surveillance, Geospatial Mapping, Computer Vision, Transformer-Based Models, GPS Geotagging.},
        month = {April},
        }