3dEYE’s cloud analytics engine processes events sent from the camera and performs object detection and tracking based on the rules configured in the Analytics section of the camera properties in the Admin Portal.
Below is an overview of how the system analyzes events and generates detections.
1. Event Intake from the Camera
The server continuously listens for events generated by the camera, including:
- Motion detection
- Line crossing
- Intrusion
- Any other configured camera-side triggers
When an event arrives, the server checks whether it matches any Detection Rules configured for that camera.
2. Rule Matching
If the incoming event matches a configured trigger:
- The server begins frame-by-frame analysis for the duration of the event.
- Each frame is scanned for the objects defined in the detection rule (e.g., person, vehicle).
3. Object Detection and Tracking
During analysis:
- Every detected object is assigned a unique tracking ID.
- The object is tracked until it leaves the camera’s field of view.
- Only new and unique objects are reported.
- The same object will not be reported again during a continuous event unless it becomes fully obstructed and then reappears.
This prevents duplicate detections and ensures clean, reliable event reporting.
4. Important Configuration Notes
- Each camera supports up to 6 analytics rules.
- Each rule can contain up to 10 analytical masks.
- Masks may overlap or be nested.
- Only objects that are at least 30% inside an analytical mask will be reported.
Proper mask placement and sizing are essential for accurate detection.
Detection accuracy depends heavily on image quality. For guidance on optimizing camera parameters (exposure, shutter speed, IR, angle, etc.), please refer to the Image Optimization article↗️.