Smart ATM Security System using AI (YOLO)

Automated Teller Machines (ATMs) are frequent targets of robbery, vandalism and arson, yet most ATM booths are protected only by CCTV cameras that record video without understanding it. This paper presents a software-only Smart ATM Security System that analyses the booth camera feed in real time with a YOLO (You Only Look Once) object detector trained on five classes: person, weapon, fire, smoke and fight. A sliding-window rule confirms a class only when it appears in K of the last N frames, and confirmed detections are mapped to three threat levels: Normal, Minor Threat and Major Threat. Every threat produces an annotated snapshot and an alert record in MongoDB, which a FastAPI backend pushes to a React dashboard over WebSocket. Published in Research Trends International Journal of Technology and Innovation (RTIJTI), ISSN 3144-3273, Volume 1, Issue 4, October – December 2026, pp. 62-66.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24375318
Primary Topic
Video Surveillance and Tracking Methods
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Smart ATM Security System using AI (YOLO)

S. D. Satre, Ashok Shirsat, Aditya Ghun, Akshay Shirsat et al.
Zenodo (CERN European Organization for Nuclear Research)
Video Surveillance and Tracking Methods
article

Smart ATM Security System using AI (YOLO)

S. D. Satre, Ashok Shirsat, Aditya Ghun, Akshay Shirsat, Revan Girawale
article en

Abstract

Automated Teller Machines (ATMs) are frequent targets of robbery, vandalism and arson, yet most ATM booths are protected only by CCTV cameras that record video without understanding it. This paper presents a software-only Smart ATM Security System that analyses the booth camera feed in real time with a YOLO (You Only Look Once) object detector trained on five classes: person, weapon, fire, smoke and fight. A sliding-window rule confirms a class only when it appears in K of the last N frames, and confirmed detections are mapped to three threat levels: Normal, Minor Threat and Major Threat. Every threat produces an annotated snapshot and an alert record in MongoDB, which a FastAPI backend pushes to a React dashboard over WebSocket. Published in Research Trends International Journal of Technology and Innovation (RTIJTI), ISSN 3144-3273, Volume 1, Issue 4, October – December 2026, pp. 62-66.

Zenodo (CERN European Organization for Nuclear Research)
Video Surveillance and Tracking Methods
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon