Stride
Pedestrian traffic flow monitoring and analysis — vision and sensor fusion counts movement patterns for civic planning.
Problem
Walkability interventions need counts, not anecdotes. Stride prototypes pedestrian flow monitoring that combines camera analytics with floor sensors to quantify peak hours, crossing behavior, and congestion hotspots.
What it covers
- Computer-vision people counting with region-of-interest masks
- Supplemental floor sensors for occlusion-heavy crossings
- Aggregated hourly flow metrics for planners
- Anonymized heatmaps without retaining raw frames
def count_crossing(frame: np.ndarray, roi: Rect) -> int:
masked = crop(frame, roi)
detections = detector.predict(masked)
return sum(1 for d in detections if d.class_id == "person")