Planograms — AI-Powered Shelf Monitoring
Real-time, zero-config shelf compliance across every camera, every store, every minute. Automatic shelf detection, continuous compliance scoring, and a complete audit trail — without weeks of manual calibration.
What Planogram does
Braingine’s Planogram module verifies product placement against your stored planogram templates automatically. It tells you, store by store and rack by rack, what’s there, what’s missing, what’s misplaced, and whether the shelf actually matches what your category managers approved — with 0-100% compliance scoring that updates as fast as your cameras can capture frames.
Key capabilities
- Automatic shelf detection — k-means clustering identifies shelf levels without manual ROI configuration; adapts to varying angles and shelf heights
- 0-100% compliance scoring with configurable confidence thresholds (default 0.60 detection, 0.75 compliance)
- 6 alert types — low compliance, missing item, wrong item, extra item, rack empty, restocking needed
- Multi-rack, multi-camera — single camera at 15-30 FPS, multi-camera at 5-10 FPS each, parallelized
- 100+ items per planogram with exact rack / shelf / X-Y positions via junction tables
- Time-series snapshots — 6+ months of detection history for trend and “missing for X hours” queries
- Full audit trail — every planogram change tracked through draft → approved → rejected → archived states
- 52 pre-built SQL reports — compliance dashboards, inventory, brand performance, regional rollups, executive KPIs
How it works
Shelf-facing camera streams feed into MaskRCNN instance segmentation or YOLOv8 at 50-150ms per frame (CPU, 2-4 cores). Detected bounding boxes are clustered into shelves automatically — no manual ROI setup. The detected layout is compared against the planogram template, producing per-shelf and overall compliance scores. Results are persisted to PostgreSQL with full audit trails, time-series snapshots, and alert triggers. Models run via OpenVINO with FP32 / FP16 / INT8 precision options.
Where it pays off
- Retail stores — out-of-stock detection, refrigerated case compliance, promotional display verification
- Warehouses — picking-face accuracy, misplaced inventory, stock-rotation enforcement
- Vending machines — availability checks and restocking alerts
- Pharmaceutical displays — regulated-substance placement verification
Why teams pick Braingine for planogram
- Auto shelf detection — no manual calibration, no per-camera ROI setup (the bottleneck of most competitor tools)
- Complete temporal history — time-series snapshots enable “missing for X hours” and trend-based analytics, not just point-in-time checks
- Governance built in — approver workflows and audit trail for regulated retail
- Multi-model — MaskRCNN, YOLOv8, or custom SKU-specific models via a central registry
- Camera calibration — perspective correction and angle compensation per display, not per software install
- 52 production-ready reports — no custom dashboarding project required
Deployed in 5 days
Like every Braingine product, the Planogram module is delivered on the 5-Day Implementation framework. From cameras to compliance scores in a working week — no six-month integration, no PhD required.