NexVDO Solutions · Smart city

Roadside video pipelines on edge compute.

Deploy multi-camera ingest and detection without shipping video to someone else's cloud.

Request an evaluation Explore VideoAI

NexVDO Solutions · Smart city

Why edge for smart city

Keep camera traffic local

Multi-camera RTSP and ONVIF ingest stays on the edge system. Cloud transport is not the only path for live feeds.

Analyze where video arrives

Flow monitoring and incident or scene detection run at the edge. The pipeline does not depend on cloud-only analysis.

Send events upstream

Edge-to-cloud event metadata gives downstream systems the event context without making every live feed the default payload.

Proven patterns

Start with the camera job. Pair the modules with real hardware.

Start with the camera job. Pair the required SDK modules with real YUAN hardware.

Multi-camera RTSP and ONVIF ingest

Bring multiple network-camera feeds into one edge pipeline for local processing and output.

SDK modules

CaptureStream

Recommended hardware

VPP6N0-S AGXJetson AGX Orin · 275/200 TOPS
EDG6N0-S T5XJetson Thor T5000/T4000 · 2070/1200 FP4 TFLOPS

Flow monitoring

Apply pedestrian and vehicle flow monitoring at the edge and emit event metadata for downstream systems.

SDK modules

AnalysisStream

Recommended hardware

VPP6N0-S AGXJetson AGX Orin · 275/200 TOPS

Incident and scene detection

Run scene recognition, object detection, tracking, and Object-Appear-in-Area against live feeds.

SDK modules

Analysis

Recommended hardware

EDG6N0-S T5XJetson Thor T5000/T4000 · 2070/1200 FP4 TFLOPS
VPP6N0-S AGXJetson AGX Orin · 275/200 TOPS

Audio backchannel for voice warning

Return a voice warning from the client to the edge endpoint through Audio Backchannel over ONVIF or RTSP.

SDK modules

Stream

Recommended hardware

VPP6N0-S AGXJetson AGX Orin · 275/200 TOPS

Edge-to-cloud event metadata

Package detection events as metadata for cloud delivery while the local edge pipeline keeps the camera workflow running.

SDK modules

AnalysisStream

Recommended hardware

VPP6N0-S AGXJetson AGX Orin · 275/200 TOPS
EDG6N0-S T5XJetson Thor T5000/T4000 · 2070/1200 FP4 TFLOPS

Capture → process → output

Keep the video path at the roadside.

IP cameras send RTSP or ONVIF feeds to YUAN Edge hardware. NexVDO Capture and Stream establish the inputs. Analysis runs flow monitoring, scene recognition, object detection, and tracking. The edge system sends event metadata and selected stream output to cloud services. Audio Backchannel returns voice warnings over ONVIF or RTSP.

Published ingest, analysis, audio, and hardware facts for a roadside video pipeline. Unpublished detail remains marked.

Pipeline job	Published path	Hardware	SDK module	Status
Camera ingest	RTSP, ONVIF	VPP6N0-S AGX / EDG6N0-S T5X	Capture + Stream	Published
Flow monitoring	Pedestrian and vehicle flow	VPP6N0-S AGX	Analysis	Published
Incident and scene detection	Object detection, tracking, Object-Appear-in-Area	EDG6N0-S T5X	Analysis	Published
Voice warning	Audio Backchannel over ONVIF and RTSP	VPP6N0-S AGX	Stream	Published
Edge-to-cloud output	Event metadata, cloud streaming	VPP6N0-S AGX / EDG6N0-S T5X	Analysis + Stream	Published
Protocol detail	RTSP and ONVIF versions	YUAN Edge hardware	Stream	Protocol versions — pending publication

YUAN Edge hardware

Every pattern runs on named hardware

The systems and capture families above come from the YUAN Edge lineup — review them together before your evaluation.

See YUAN Edge hardware

Build resources

Published material stays with the supported evaluation path. Missing artifacts stay visible as honest states.

Sample pairings
Capture hardware, edge systems, and SDK modules arranged by pipeline job.
published with evaluation access
Reference architecture
A signal-to-output architecture for this vertical.
in preparation
Developer hub
Module guides, support routes, and integration notes.
Go to the developer hub