Tracker

Overview

Tracker node assigns persistent track IDs to rectangular detections across frames and returns tracked objects with history, counts, and optional overlays.

Connect detections from nodes such as General Object Detection, Object Detection, or People Detection. Downstream nodes such as Tracker Line Count, Tracker Zones Count, and Tracker Heatmap consume the tracked objects output.

Input

Input Image

image required

The image frame used for tracking visualization. Connect this to a camera or upstream image output.

Rectangular Objects

array required

List of rectangular detection objects from an upstream node. Each object must include at least a Bounding Box (bbox: [x, y, width, height]). Label, Confidence, and Extra Data use defaults when missing.

Expected Frame Interval (ms)

integer

Expected interval between frames in milliseconds.

Default: 33

Labels to track (filter)

array required advanced

If empty, all labels are tracked. When set, only detections whose labels are in this list are tracked.

Default: []

Tracking Threshold

number advanced

Threshold for tracking confidence.

Range: 0.0 to 1.0
Default: 0.23

Minimum Track Frames

integer advanced

Minimum number of frames required to establish a track.

Range: 1 to 20
Default: 3

Maximum Track Lost Frames

integer advanced

Maximum number of frames to keep a lost track before dropping it.

Range: 1 to 1000
Default: 30

Matching Threshold

number advanced

Threshold for matching new detections to existing tracks.

Range: 0.0 to 1.0
Default: 0.8

Maximum Keep Frames

integer advanced

Maximum number of frames to keep a track without updates.

Range: 1 to 100000
Default: 500

Track Smooth Alpha

number advanced

Smoothing factor for track position updates.

Range: 0.0 to 1.0
Default: 0

Average Object Attributes

boolean advanced

When enabled, averages each tracked object's attributes over its history and exposes them as avg<AttributeName> on the tracked object. Mean is used for integer/float, mode for string/boolean; other types and timestamp are ignored. Empty or unknown strings and zero values are skipped.

Default: true

When enabled, Override Numerical Attribute Average Methods can appear: Attribute Name, Average Method (mean, median, mode, max, min, sum), and Ignore Zero Values.

Overlay results

boolean required advanced

Whether to draw tracking results on the output frame. See Overlay Results.

Default: true

When enabled, advanced overlay options can include:

• Draw Point (default true)
• Draw IDs (default true)
• Draw Bounding Boxes (default false)
• Draw History Trails (default true)
• Point Position — TOP, MIDDLE (default), BOTTOM
• Visualize Scale (default 1, range 0–10)
• Show Lost and New (default false)
• Smoothing Alpha (default 0, range 0.0–1.0)

Output

Output Image

image

Output frame with track overlays when overlay is enabled.

Track Counts

object

Summary counts for the current frame:

• Total
• Active
• Lost
• New

Tracked Objects

array

Array of tracked objects. Each object contains:

• Track ID
• Object History — per-frame Center Point, Bounding Box, Label, Confidence, Extra Data, Timestamp (Unix ms)
• Frames Since Last Update
• Tracking Time (ms)
• Optional averaged attributes (avg…) when Average Object Attributes is enabled