Technology
A deep look at the computer vision, hardware, and software systems that power BusGuard’s real-time detection capabilities.
Violation → alert
Incident logged
IllustrativeWritten to 256 GB onboard storage
01Detection0 ms
Computer vision continuously monitors the stop arm and surrounding traffic. The moment the arm extends and a vehicle crosses the threshold, analysis begins.
02Analysis~80 ms
The on-device CNN scores the frame, filtering parked cars, cyclists, and shadows. Only confirmed violations pass the 94% confidence threshold.
03Alert<200 ms
An audible 100 dB alert fires through the onboard buzzers. Simultaneously, a timestamped record with GPS, image capture, and route ID is written to storage.
Incident logged
IllustrativeWritten to 256 GB onboard storage
Core intelligence
Our proprietary CNN is trained specifically on stop-arm violation scenarios, accounting for bus geometry, arm extension angles, vehicle trajectories, and lighting variation. 95.3% true positive rate across 211 field trials.
Live detection feed
IllustrativeAudit trail
Every detected violation is timestamped and stored with full metadata, GPS coordinates, image capture, vehicle speed estimate, and route ID, creating a legally defensible audit trail for district reporting.
Incident log
IllustrativeHardware
A compact, weatherproof unit that mounts inside any school bus and runs on the bus’s own 12V power.
Detection
95.3%
True positive rate
Across 211 trials. Day, night, rain, or shine.
Custom CNN · proprietary dataset
Latency
<200ms
Start to buzzer fired
Alert
100 dB
Onboard buzzer
Storage
256 GB
Onboard, expandable
Enclosure
IP67
Weatherproof
Camera
1080p
Wide-angle, WDR
Connectivity
4G LTE
+ WiFi sync
Power
12V
Bus electrical
Operating temp
−20°C to 60°C
False positives
<2%
Parked cars, cyclists, and shadows filtered out
T+200 ms · Alert fired
Contact sales, ask about pricing, or discuss a district deployment. We respond within one business day.
