A&A
Smart RPC Real-time Grain Inventory System
An IoT system that measures grain stock inside rice processing complex (RPC) silos automatically and manages the equipment remotely. First built in 2023 and extended year by year, it now runs across 18 RPC sites and 103 silos in Korea.
Aug 2023 – ongoing
Project
- Aug–Dec 2023 — built the Smart RPC real-time inventory management system.
- Jan–Dec 2025 — delivered the enhancement project for the same system.
- Sensors installed in the silos measure the stored grain automatically, and equipment across multiple regions is monitored and managed remotely.
- We built every part ourselves: the sensor hardware, the collection program on the edge device, the server, the dashboard and the mobile app.
Challenge
- Stock levels inside a silo were checked by eye and by hand, so nobody knew the current figure.
- A grain surface is not flat — measuring a single point does not tell you how much is stored.
- Equipment is spread across the country, so every fault had to be investigated site by site.
- Inspecting equipment or changing a setting required someone to travel to the site each time.
Architecture
We designed and implemented the whole path: readings from the field sensors pass through an embedded board into the IoT server, and from there into the dashboard and the customer’s ERP.
- Silo sensorsLiDAR range sensor, temperature and humidity sensor
- Grain flow meterWeight sensor
- Edge deviceCollects and pre-processes readings, talks to the collection server
- IoT serverDatabase, API server, ingestion and processing
- Dashboard / ERPMonitoring, history, anomaly alerts, ERP integration
Development
Sensor node and hardware
- Developed the LiDAR sensor hardware in house — 40 m range, IP65 rated.
- The LiDAR range sensor measures the distance to the grain surface at fine intervals.
- A pan-tilt module driven by two servo motors scans the whole silo interior rather than a single point.
- A temperature and humidity sensor records the storage conditions alongside it.
- A Raspberry Pi 4B edge device runs the collection program, pre-processes the readings and posts them to the API.
Inventory calculation from point clouds
- The scan is reconstructed into a 3D point cloud.
- Least-squares circle fitting detects the silo wall and separates wall points from interior points.
- Obstacles are distinguished from the grain surface.
- Shadowed areas the sensor could not reach are detected and filled in by pattern matching.
- Analysis and 3D visualisation are built with scipy, scikit-learn and Open3D.
Dashboard
- Per-region and per-silo monitoring — stored weight, temperature, humidity and history.
- Grain flow meter monitoring — husking, whitening and milling yield.
- Alerts to the administrator when an anomaly appears.
- Filtering by period, region and silo, with Excel export.
- Works on mobile.
2025 enhancement
- A single monitoring page across all RPC sites, showing every fault in one view.
- An RPC records page — equipment and silo details, server list, user and group permissions.
- A VMS management app that connects to field equipment over Bluetooth to diagnose, configure and control it on the spot, with no site visit.
- Device settings managed and pushed as YAML.
- Silo capacity and sensor height calculators.
Screens
These are the live product screens. The interface is in Korean.
Result
The system has been extended every year since the first installation in 2023.
| Phase | Year | RPC sites | Silos |
|---|---|---|---|
| Phase 1 | 2023 | 1 site | 49 silos |
| Phase 2 | 2024 | 4 sites | 29 silos |
| Phase 3 | 2025 | 10 sites (pilot) | 10 silos |
| Phase 3 | 2025 | 3 sites (new) | 15 silos |
| Total | 2023–2025 | 18 sites | 103 silos |
Technology
Related links
Other projects
Korea Press Foundation
2021–2025
KPF-BERT and AI-based News Services
Korea Data Agency
2022 – ongoing
AI Data Processing for the Data Voucher Programme
Codnut
2023–2025