Ikshana Video Analytics
25+ use cases. Live across India's most demanding public safety environments.
-
AI and Deep Learning Analytics Engine
At the core of Ikshana is a deep learning engine trained on real Indian operational datasets — not generic global models. It understands the specific visual environment of Indian cities, transport systems, border areas, and institutional spaces — delivering higher detection accuracy and lower false-alarm rates than off-the-shelf international platforms.
-
Real-Time Event Detection and Classification
Every detected event is classified — not just flagged. The system tells operators what happened, where it happened, which camera captured it, and what the confidence level of the detection is. Operators receive intelligence, not alerts that need manual investigation to understand.
-
Edge-to-Cloud Architecture
Analytics runs at the edge — on local processing units at the camera site — generating alerts locally without central server dependency. All events are simultaneously transmitted to the central platform for aggregation, search, and command-level visibility. Monitoring continues even when central connectivity is degraded.
-
Verified Alert Delivery
Alerts are delivered via dashboard notification, SMS, WhatsApp, and email — with event image and video clip attached. Every alert is verified by the AI before it reaches an operator — eliminating the false-alarm volume that makes manual monitoring systems operationally unsustainable.
Use Cases by Category
- People counting and crowd density estimation
- Crowd anomaly detection — unusual gathering, sudden dispersal, aggressive movement
- Face recognition — watchlist matching for wanted, suspicious, or VIP individuals
- Crowd heatmaps and movement flow analysis
- Tailgating and piggybacking at access-controlled entry points
- Dwell time analysis — prolonged presence in sensitive areas
- Person fall and collapse detection for medical emergency response
- Perimeter intrusion and trespassing
- Fence jumping and boundary breach
- Loitering in sensitive or restricted areas
- Unattended object detection in public spaces
- Fire and smoke detection from camera feeds
- Camera tampering — covering, repositioning, or spray-painting
- Unauthorised vehicle access in restricted zones
- ANPR — Automatic Number Plate Recognition with OCR
- Watchlist matching against VAHAN, SARATHI, and custom categories
- Red light violation detection
- Overspeed detection
- Wrong-way driving
- Triple riding detection
- No helmet detection
- No seatbelt detection
- Wrong parking and stopped vehicle detection
- Traffic congestion and density analysis
- Automatic e-challan generation with ITMS integration
- Stray animal detection on roads
- Garbage overflow and cleanliness monitoring
- Pothole and road condition detection
- Equipment malfunction at infrastructure sites
- Worker behaviour monitoring at operational sites
- Long-range human, vehicle, and drone classification
- Thermal intrusion detection in forward areas
- FOD detection on airfield runways
- Perimeter breach at defence installations
- Naval perimeter and waterborne threat detection
- Restricted zone access monitoring
Sectors and Deployments
- Public Safety - Chennai Mega City Surveillance · MTC Transit Safety Network
- Critical Infrastructure - FCI — 561 depot surveillance network
- Traffic Enforcement - TROZ — Traffic Regulatory Zone, Chennai
- Defence - Indian Army border surveillance — 79 forward area systems
- Education - 22,931 school camera network — Tamil Nadu and Punjab
- Naval - Naval base perimeter and water monitoring
- Airbase - FOD detection and airfield surveillance — IAF
Technical Capabilities
- Deep learning AI engine — trained on Indian operational datasets
- 25+ detection and classification use cases
- Real-time edge processing — local alert generation without central dependency
- Multi-feed simultaneous analysis — hundreds of cameras concurrently
- ANPR with OCR — plate recognition and watchlist matching
- Face recognition — watchlist and VIP matching
- Thermal camera analytics — night and all-weather detection
- Verified alert delivery — dashboard, SMS, WhatsApp, email with image and video
- Event search — by camera, location, event type, date, time, plate number
- Evidence export — timestamped, camera-attributed, legally defensible
- API integration — ITMS, e-challan, VAHAN, SARATHI, command centre platforms
- Edge-to-cloud architecture — continuous operation in degraded connectivity
Deployment Models
- Edge Deployment — Analytics runs on local processing units at each camera site. Ideal for remote locations, bandwidth-constrained environments, and defence forward areas.Centralised Server Deployment — Analytics runs on central servers with camera feeds streamed in. Ideal for high-bandwidth urban environments and command centre integrations.Cloud Deployment — Fully cloud-hosted analytics for organisations without on-premise server infrastructure.Hybrid — Edge processing for real-time local alerts combined with cloud aggregation for central command visibility and cross-site analytics.
Customisation Options
- Detection use case selection — configure exactly which analytics run on which cameras
- Alert thresholds — sensitivity configuration per use case per camera
- Watchlist management — custom categories beyond standard presets
- Alert routing — define who receives which alert type by role and location
- Integration — any command centre, dispatch, or enforcement platform via open API
- Camera compatibility — integrates with any IP camera brand or existing infrastructure
- Reporting — custom report formats for compliance, audit, and operational review