waddleloo

Waterloo’s AI goose tracker

AI Analysis

Report a Sighting

Upload a photo of geese on campus. YOLO detects and counts them. Gemini Vision checks for goslings, nesting behaviour, and aggression signals. Approved sightings feed into the live campus map.

Drop a photo here

JPG, PNG · Max 20 MB

35%

Lower = more detections, higher = fewer but more certain

What happens next

01

YOLO detection

A YOLO-based model scans the frame, draws bounding boxes around every goose detected, and returns a count with per-detection confidence scores.

02

Severity assessment

Goose count and campus location context are combined to assign a severity level: low, moderate, or high. High counts near known nesting zones escalate automatically.

03

Visual intelligence

Gemini Vision

A Gemini Vision pass analyses the frame for goslings, nesting posture, spread wings, and lowered-neck aggression signals, the cues that matter most for safety.

04

Human review

Submitted sightings go through a quick review before going live. This keeps the map trustworthy and filters out misidentified birds or duplicate reports.

05

Live campus hotspot

Approved sightings appear as real-time markers on the campus map, visible to anyone checking routes before they walk.