SolutionEngine LogoSolutionEngine
HomeProductUse CasesPricing

Datasource: RTSP Stream

Use RTSP Stream for camera-driven computer vision pipelines.

This datasource opens an RTSP stream and emits sampled frames as events at controlled FPS and resolution.

Best Use Cases

  • Live defect detection
  • Security and monitoring analytics
  • Edge vision workflows with controlled sampling

Configuration Fields

FieldRequiredTypeDefaultPurpose
urlYesstring-RTSP stream URL
authTypeNoselectnoneAuth mode: none, basic, digest
usernameConditionalstring-Required when auth is basic or digest
passwordConditionalpassword-Required when auth is basic or digest
fpsNonumber5Frame sampling rate (1-8)
resolutionNoselect640x480Output frame resolution

How to Configure Correctly

  1. Select RTSP Stream datasource type.
  2. Set valid camera URL (rtsp://host:port/path).
  3. Configure auth fields only if camera requires authentication.
  4. Start with fps=3-5 for baseline throughput.
  5. Choose resolution based on model requirements and compute budget.
  6. Save and run a short validation workflow.

Example Config

{
  "url": "rtsp://camera-user:camera-pass@192.168.1.45:554/stream1",
  "authType": "basic",
  "username": "camera-user",
  "password": "<secret>",
  "fps": 5,
  "resolution": "640x480"
}

Performance Guidance

  • Lower FPS reduces CPU/GPU load and storage cost
  • Match resolution to model training/inference input expectations
  • Use edge environments when camera latency is critical

Common Problems

  • Stream cannot open: invalid URL/path or network reachability
  • Black frames/timeouts: camera instability or bitrate constraints
  • High latency: FPS/resolution too high for available resources

Related Pages

  • Continue with Datasource - File Stream for replay and testing workflows.
  • Continue with Workflows for vision pipeline design.
PreviousDatasource - MQTT Subscribe
NextDatasource - File Stream