EventDetection
Configuration for event detection enrichment on video time-series data.
Event detection generates embeddings from video data enabling natural language similarity search on events. This allows customers to:
-
Query video events using semantic search after enrichment completes
-
Find relevant video segments through natural language queries
-
Search across video time-series data stored in IoT SiteWise
You must specify the dataset, exactly one time-series identifier (timeSeriesId OR propertyAlias), and trim settings defining the video time window to process.
Contents
- datasetId
-
The IoT SiteWise dataset ID containing the video time-series data to analyze. Query IoT SiteWise to discover available datasets in your workspace.
Type: String
Length Constraints: Fixed length of 36.
Pattern:
^(?!00000000-0000-0000-0000-000000000000)[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$Required: Yes
- trimSettings
-
Time range settings defining which portion of the video time-series data to process. Required to ensure predictable processing time and prevent analyzing unbounded datasets. Start and end times must be within the dataset's time bounds.
Type: EnrichmentTrimSettings object
Required: Yes
- propertyAlias
-
Human-readable alias for the video time series to analyze (e.g., /camera/warehouse/zone-a). Specify either propertyAlias or timeSeriesId, but not both. Use this when you have configured friendly aliases in IoT SiteWise for better readability.
Type: String
Length Constraints: Minimum length of 1. Maximum length of 2048.
Pattern:
[^\u0000-\u001F\u007F]+Required: No
- timeSeriesId
-
Unique system identifier for the video time series to analyze. Specify either timeSeriesId or propertyAlias, but not both. Use this when you have the system-generated time series identifier from IoT SiteWise.
Type: String
Length Constraints: Minimum length of 36. Maximum length of 73.
Required: No
See Also
For more information about using this API in one of the language-specific Amazon SDKs, see the following: