Ai Crowd Monitoring System for Mahalaxmi Mandir Kolhapur
Manual CCTV couldn't proactively manage stampede risk or identify threats during massive festival crowds. AI heatmaps, facial recognition, and vehicle tracking now enable real-time crowd safety and rapid response.
Problem Statement
Mahalaxmi Mandir witnesses massive devotee crowds during festivals and daily darshan, making visitor movement, entry/exit management, and public safety difficult to monitor manually. Overcrowding and blocked pathways raise stampede risks, while traditional CCTV only enables reactive, manual monitoring delaying detection of suspicious activity and identification of criminals or missing persons in dense crowds.
Core Gap: Authorities need a centralized, real-time AI platform for crowd management, security, and rapid emergency response during high-footfall periods.
Solution
Implemented in and around Mahalaxmi Mandir to ensure devotee safety inside temple premises and smooth traffic flow in surrounding high-density zones, using AI-powered heatmaps, facial recognition, vehicle counting, and number plate recognition for real-time monitoring and proactive alerts.
By strengthening coordination across police, traffic, disaster management, and urban development departments, it stands as a flagship Smart City and Safe Pilgrimage model for replication at other pilgrimage sites.
Functionalities
Facial Recognition (FR)
Heat Maps
Tech Stack
Layer
Technology Stack
Computer Vision
Python, YOLOv11, PyTorch, OpenCV, CUDA, NVIDIA GPU
Dashboard
React.js, Node.js,FastAPI, WebSockets, Mapbox/Leaflet, Chart.js
Reporting
Python, Pandas, SQL, ReportLab, Excel (OpenPyXL), Power BI
Hardware
NVIDIA RTX/A-Series GPUs, Ubuntu Linux, Docker, NVIDIA CUDA Toolkit, Gigabit Networking
Impact on Client
Public safety & trust
2–3 missing persons were located and unauthorized entries were prevented, directly boosting devotee safety and confidence during the event.
Faster incident response
Real-time AI alerts (crowd density, FR matches) enabled security and police to act on-site within minutes instead of relying on manual monitoring.
Risk mitigation
Early detection of high crowd density (81% near the CCTV room) and 3–4 flagged suspicious persons helped prevent potential crowd-crush or security incidents before they escalated.
Operational visibility
Gate-wise visitor trend data (East Gate busiest, North/South/West spikes on Sep 28) gives the client actionable insight to plan staffing, crowd redirection, and infrastructure (e.g., more cameras) for future events.
Identified gaps for improvement
The single-camera blind spot near the CCTV room is a clear, low-cost fix the client can act on to close a current surveillance gap.
Net effect
The system moved the client from reactive to proactive crowd and security management — improving safety outcomes, response time, and planning accuracy for large gatherings, while surfacing a concrete infrastructure gap to fix next.
Testimonials
"During the high-footfall Navratri festival at Mahalaxmi Temple, this AI-powered Crowd Management System was a game-changer. It provided live visibility into bottleneck areas across all access gates while empowering Kolhapur Police with facial recognition to track known offenders from police databases. It’s a remarkable innovation in public safety and smart administration."
Amol Yedge
Former Collector & District Magistrate,Kolhapur
ZP
September 2025
"During the high-footfall festivals at Mahalaxmi Temple and Jotiba Temple, managing large crowds across multiple access points was a major challenge. The AI-powered Crowd Management System gave the Devsthan Samiti real-time visibility into crowd movement, helping us identify bottlenecks and take timely action. It has significantly strengthened our ability to manage pilgrim movement safely and efficiently."
Shivraj Naikwade
Secretary
ZP
