Detection Configuration
Detection rules applied within a region.
Example
{
"class_name": "person",
"class_threshold": 10,
"object_size": 150,
"combined_threshold": 80,
"object_descriptions": [
{
"text": "person carrying a box",
"threshold": 90,
"alert": true,
"background_prompt": false
}
],
"face_recognition": {
"match_type": "group",
"match_threshold": 80,
"group": "employees"
},
"similarity": {
"images": []
},
"conditions": [
{
"type": "count_greater_than",
"count": 0
}
],
"alert_seconds": 1,
"reset_seconds": 5
}
Fields
| Field | Type | Description |
|---|---|---|
class_name | string | Optional class filter such as person, vehicle, bag, animal, or unknown. |
class_threshold | number | Confidence threshold for class_name. |
object_size | number | Minimum object size in pixels. |
combined_threshold | number | Minimum combined threshold used by the server when alerting. |
min_contour_area | number | Optional contour-area filter for motion or extraction logic. |
max_bounding_box_area | number | Optional upper bound for bounding-box area. |
object_descriptions | Object Description[] | Natural-language prompts used for matching and alerting. |
face_recognition | Face Recognition Config | Face recognition match settings. |
similarity | Similarity Config | Similarity-search settings. |
conditions | Detection Condition[] | Alert conditions such as count rules or line crossing. |
alert_seconds | number | Time a match must persist before the server alerts. |
reset_seconds | number | Time without a match before the alert state resets. |
objects | Detection Object[] | Optional advanced object filters or labels. |