Out-of-stock and in-store traffic
Empty-shelf detection, visitor counting, checkout queue control, and loss prevention — using cameras you already have.
Computer vision, industrial monitoring, and custom models running on-device. No sending data to the cloud. No waiting seconds for a response. For companies that need AI to react before the problem is even visible.
Peak of 5.2 alf/sec at 2pm. 3 defective alfajores discarded and line 2 was stopped for 8 min — the system alerted the supervisor over WhatsApp.
YOLO + ROIs detect and count every alfajor that moves down the line. From there come live metrics — alfajores per minute, idle time — and alerts when the line stops. On-device at the factory, with nothing sent to the cloud.
One general ROI + one ROI per spot. The model decides whether there's a vehicle or not in each slot, and the dashboard up top shows the open spots live. It works with the camera that was already installed.
Four places where cloud latency isn't enough — and where we know how to put custom models running on your hardware.
Empty-shelf detection, visitor counting, checkout queue control, and loss prevention — using cameras you already have.
Defect detection, poorly sealed packages, crooked labels — with automatic line stop when the model calls it.
Pallets, boxes, and containers tracked across multiple cameras. OCR / QR reading on the move. Reconciliation with your WMS.
Animal weight estimation, head counting, perimeter monitoring with drones, NDVI over fields — without depending on the internet.
Send us an email with three lines — what problem you want to solve, where, and what scale you're handling. We'll get back to you with a yes, a no, or a question. We sign an NDA before sharing any sensitive info.
edge@carpin.ai