Retrofitting an Old Parking Lot with IoT: A Retrospective from Manual Registration to Seamless Payment

The client was a surface parking lot in an old city district: 300 spaces, two entrances/exits, and four cashiers working in three shifts. Their ask was straightforward—reduce manual labor and plug the leaks from missed and evaded fees. This is our retrospective of the whole process, from the initial site survey to final acceptance.

The Debate: Geomagnetic Sensors or Video Poles

• Geomagnetic detectors: buried in the center of each space, they detect magnetic field changes to determine whether a vehicle is present. Pros: cheap, fast to install, weatherproof. Cons: they cannot read license plates, so they must be paired with handheld terminals or QR-code payment.
• Video poles: one camera per space, with license plate recognition done in a single step—but the unit price is five times that of a geomagnetic sensor, and tree shade, standing water, and nighttime glare all drag down recognition accuracy.

In the end we chose a hybrid approach: license plate recognition cameras at the entrances/exits (the backbone of seamless payment), and geomagnetic sensors at the space level for occupancy statistics. Geomagnetic data is reported to the platform in real time, and the guidance screen at the entrance shows remaining spaces—this alone cut peak-hour entrance queues in half.

Networking: Why 4G Instead of Fiber

The parking lot had no telecom riser, and the application and road-excavation approvals for bringing in fiber would have taken two months. We switched to a 4G Cat.1 industrial gateway: monthly traffic is under 2GB, the cost is a fraction of fiber construction, and the deployment timeline shrank from two months to three days. In retrofit projects, "good enough and fast" often beats "optimal."

Offline Fallback: No Network Must Not Mean No Charging

No matter how reliable the platform is, carrier networks have their shaky moments. We set three hard rules for the lot:

1. The barrier gates at entrances/exits cache the whitelist and the last 24 hours of transaction records locally, so they keep lifting as usual when the network is down;
2. Payment QR codes support "leave first, pay later," so drivers never have to wait at the exit for the network to recover;
3. Once the network recovers, offline data is backfilled automatically, and the platform's reconciliation module uses the local transaction records as the source of truth for discrepancy checks.

Six months after going live, the lot has weathered two carrier outages with zero driver-side impact.

The Numbers After Going Live

• Cashiers reduced from 4 to 1 (handling exceptions and inspections);
• Monthly lost revenue from fee evasion dropped from roughly 8% to under 1%;
• Space turnover rate increased by 17%—the guidance screen helps drivers find spaces faster, and leave faster too.

Retrospective

The hard part of these brownfield retrofit projects is never the platform's features—it's fitting in with the old workflows: the cashiers' habits, the drivers' expectations, the property manager's way of reconciling accounts. Technology is the lever, but the fulcrum is always the site itself.

Related Documentation

• Video Surveillance: live streaming, playback, and PTZ control after the entrance/exit license plate recognition cameras are connected to the platform via GB/T 28181.
• Alert Center: alert notification configuration for events such as network outages and device anomalies, supporting the offline fallback strategy described in this article.
• Device Management: onboarding, grouping, and status monitoring for devices such as geomagnetic detectors and barrier gates.