Skip to content

scenescape-integration.md

Latest commit

 

History

History
222 lines (154 loc) · 6.56 KB

scenescape-integration.md

File metadata and controls

222 lines (154 loc) · 6.56 KB

Integrate Intel® SceneScape with Smart NVR

This guide covers the integration of Intel® SceneScape with Smart NVR for enhanced traffic monitoring using live data from smart intersection application.

Overview

Smart NVR system integrates with Intel® SceneScape to enable:

  • Real-time object counting and tracking (vehicles, pedestrians)
  • Traffic flow analysis
  • Automated event routing based on count thresholds
  • Enhanced surveillance for smart intersection management

Prerequisites

Smart Intersection Reference Implementation: We will use the Smart Intersection application to showcase the integration with Intel® SceneScape.

  1. Set up the Smart Intersection application.

    # Clone smart intersection repository inside smart nvr directory if not already done
git clone https://github.com/open-edge-platform/edge-ai-suites.git -b v1.2.0

# From the Smart NVR directory, copy the DL Streamer configuration (enables RTSP streaming)
cp ./resources/si-rtsp-config.json edge-ai-suites/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/src/dlstreamer-pipeline-server/config.json

# Copy the SceneScape compose configuration
cp ./resources/compose-scenescape-rtsp.yml edge-ai-suites/metro-ai-suite/metro-vision-ai-app-recipe/compose-scenescape.yml

  2. After copying the files, start the Smart Intersection application:

    # Navigate to metro-vision-ai-app-recipe directory
cd edge-ai-suites/metro-ai-suite/metro-vision-ai-app-recipe/

# Start Smart Intersection
./install.sh smart-intersection
docker compose up -d

# Navigate back to Smart NVR
cd ../../../../smart-nvr

    The files provide:

    • RTSP streaming support in the DL Streamer pipeline,
    • Intel® SceneScape-specific Docker Compose settings.

Installation and Setup

Step 1: Get MQTT Credentials

# Get MQTT credentials from Smart Intersection
cat edge-ai-suites/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/src/secrets/browser.auth
# Expected: {"user": "<user>", "password": "<password>"}

Step 2: Configure Environment Variables

# Enable SceneScape Integration
export NVR_SCENESCAPE=true

# MQTT Configuration (from browser.auth JSON)
export SCENESCAPE_MQTT_USER="<user>"
export SCENESCAPE_MQTT_PASSWORD="<password>"
export SCENESCAPE_THROTTLE_INTERVAL=2.0  # Optional: throttle interval in seconds

Step 3: Start Smart NVR

# Start the application
./setup.sh start

# Or restart with new configuration
./setup.sh restart

Note: The setup script automatically copies Intel® SceneScape certificates from Smart Intersection if available. If certificates are missing, setup will fail with an error message.

Step 4: Verify Integration

Check logs to confirm Intel® SceneScape connection:

docker logs nvr-event-router -f
# Look for: "SceneScape MQTT client started"

User Interface Changes

With Intel® SceneScape Enabled and SceneScape Source Selected

SceneScape Enabled Interface

When Intel® SceneScape is enabled (NVR_SCENESCAPE=true) and "scenescape" source is selected:

  • Source dropdown shows both "frigate" and "scenescape" options
  • Count field becomes visible and editable
  • Users can set minimum count threshold for rule triggering (e.g., 5, 10, 15)
  • Rules table includes "Count" column for tracking thresholds
  • Count validation ensures non-negative integers only

With Intel® SceneScape Enabled but Frigate Source Selected

Frigate Selected Interface

When Intel® SceneScape is enabled but Frigate source is selected:

  • Currently Frigate object detection is disabled in this mode
  • Source dropdown still shows both "frigate" and "scenescape" options
  • Count field is automatically hidden (not applicable for Frigate)
  • Standard Frigate rule configuration with detection labels
  • Rules table shows "Count" column but displays "-" for Frigate rules
  • Full Frigate functionality remains available

Auto-Route Events Configuration

Creating Rules

Steps (both sources):

  1. Navigate to Auto-Route Events tab
  2. Select Source: "scenescape" or "frigate"
  3. Set Count: (SceneScape only) Define minimum threshold (e.g., 5)
  4. Select Camera: Choose target camera
  5. Choose Detection Label: Select object type
  6. Select Action: "Summarize" or "Add to Search"
  7. Click Add Rule

Key Differences:

  • SceneScape: Count field visible when selected
  • Frigate: Count field hidden

Rule Behavior Examples

SceneScape Rule Example:

Source: scenescape
Camera: camera1
Count: 5
Label: vehicle
Action: Summarize

Triggers video summarization when 5+ vehicles detected in camera1

Frigate Rule Example:

Source: frigate
Camera: livingroom
Label: person
Action: Add to Search

Adds person detection events to search index for livingroom camera

Troubleshooting

Common Issues

SceneScape features not visible:

# Check and set environment variable
echo $NVR_SCENESCAPE  # Should show 'true'
export NVR_SCENESCAPE=true
./setup.sh restart
# Refresh browser (Ctrl+F5)

No SceneScape events received:

# Check MQTT connection
docker logs nvr-event-router | grep -i scenescape

Debug Commands

# Check environment variables
env | grep NVR_SCENESCAPE
env | grep SCENESCAPE

# Monitor MQTT messages
docker logs nvr-event-router -f | grep "scenescape"

# Check UI logs
docker logs nvr-event-router-ui -f

# Verify SceneScape MQTT connection
docker logs nvr-event-router | grep "Scenescape MQTT client"

Monitoring Commands

# Check system CPU usage and load
cat /proc/loadavg && docker stats --no-stream --format "table {{.Name}}\t{{.CPUPerc}}\t{{.MemUsage}}"

Support

For Intel® SceneScape integration issues:

  1. Certificate Error: Ensure Smart Intersection application is running and has generated certificates
  2. Environment Variables: Verify NVR_SCENESCAPE=true and MQTT credentials are set
  3. MQTT Connection: Check logs for "Scenescape MQTT client started" message
  4. Smart Intersection: Confirm Smart Intersection application is accessible at expected path
  5. Performance Issues: Run cat /proc/loadavg && docker stats --no-stream to check CPU usage and system load
  6. High Resource Usage: If load average >10 or containers show >500% CPU, restart high-usage containers
  7. Review logs using debug commands above and contact support with relevant excerpts

For general Smart NVR issues, see the Troubleshooting Guide.