Ingest locally
RTSP, ONVIF, IP and USB feeds terminate at the edge gateway.
NviroView AI Vision Gateway
Turn existing cameras into real-time operational event sources.
NviroView reuses IP, RTSP, ONVIF and USB camera infrastructure, adds edge AI inference close to site, and produces structured events, alerts, overlays, recordings and APIs for Nvirosense workflows.
Edge-first event flow
Process locally, send events upstream and sync richer evidence when needed.
That architecture reduces bandwidth pressure, supports remote or unstable links and keeps the upstream platform focused on events, classifications, counts, alerts and reviewable evidence instead of raw continuous video wherever that is not required.
Infer locally
AI models analyse frames near site so latency and uplink demands stay lower.
Send structured events
Counts, detections, alerts, clips and snapshots can move upstream instead of every full stream.
Sync evidence
Recordings and review packs can be retained or uploaded according to the operational requirement.
Problem and opportunity
Most CCTV infrastructure records video, but leaves operational value locked inside footage.
NviroView shifts camera systems from passive recording toward proactive detection. Existing camera feeds become live views, sampled frames, AI detections, searchable metadata, evidence and alert triggers without requiring a full camera replacement. The preferred deployment model is edge-first: infer close to site, keep bandwidth under control and send structured events instead of forcing every stream upstream continuously.
Reuse
Existing camera feeds and stream protocols.
Process
Edge AI inference close to site.
Act
Events, alerts, APIs, overlays and recordings.
AI vision pipeline
RTSP ingest, stream management, frame sampling, inference, filters and output channels.
Camera ingestion
EZVIZ, Hikvision, Dahua, ONVIF-capable IP cameras, generic RTSP and local USB sources can be normalized into usable views, recordings and event metadata.
Edge architecture
Camera, edge and output layers keep inference close to site, reduce bandwidth pressure and route events to dashboards, APIs and reports.
Market shift
Traditional CCTV is reactive. Edge AI supports proactive event detection and faster response without cloud-only dependence.
Hardware fit
Choose compute by camera count, model size, resolution, environment and upgrade path.
NviroView can be positioned from low-bandwidth event nodes through RK-class pilot gateways, Hailo industrial inference, Jetson Orin modules and higher stream-count edge deployments. The public story is simple: the gateway tier should match the operational problem.
LoRaWAN event node
RK3576 AI box
Hailo-8 gateway
Jetson Orin Nano
Orin NX
AGX Orin
Deployment opportunities
Warehouses, factories, farms, security estates, retail and municipal sites.
Vision AI becomes most valuable when it is connected to a business workflow: restricted zones, PPE areas, vehicle movement, queue monitoring, after-hours activity, perimeter events, low-bandwidth remote alerts and evidence packages.
Security
Perimeter, vehicle and intrusion events.
Safety
PPE zones, hazard areas and forklift interactions.
Operations
Footfall, queue, asset and vehicle movement metadata.
Remote sites
Low-bandwidth event monitoring where continuous video is impractical.
From camera feed to business action
Ingest, process, detect, alert and preserve evidence.
NviroView upgrades camera infrastructure into an intelligent detection, recording, alerting and integration platform that can feed Nvirosense dashboards and operational workflows.
Hardware options
Deployment fit by device tier.
Match the NviroView gateway to the camera workload, bandwidth profile and model requirements before confirming final sizing.
LoRaWAN TinyML Event Camera
Event node - ~0.1 TOPS
Remote event detection only. No continuous video; sends event metadata, making it useful for farms, gates and low-bandwidth sites.
RK3576 AI Gateway
AI box - 6 TOPS
Low-cost AI entry tier for pilot deployments and smaller camera groups running light models.
Hailo-8 Industrial Gateway
R2145-12 - CM5 - 26 TOPS
Industrial Hailo inference tier for optimized HEF models and deployments that need a hardened gateway form factor.
Jetson Orin Nano Module
Jetson - up to 67 TOPS
Compact Jetson edge tier for multi-camera deployments where model choice, resolution and FPS drive sizing.
Jetson Orin NX Module
Jetson - up to 157 TOPS
Multi-stream Jetson tier for larger camera groups and heavier models, sized to the deployment workload.
Jetson AGX Orin Module
Jetson - up to 275 TOPS
High-density AI compute tier for the largest edge deployments and higher aggregate inference demand.