Counting machine output and production performance: per-machine output and efficiency with camerasCameras automatically count products coming off machines (rolls, bolts, sheets); per-machine production quantity, speed and performance are logged with AI.2024-09-23
In manufacturing the most basic question is simple, but the answer usually comes at the end of the day from manual records: which machine produced how much? Hype Vision connected the existing cameras on production and packing lines to AI: it counts the products coming off machines (rolls, bolts, sheets, etc.) one by one, tracks each, and derives per-machine production performance in real time — all fully on-premise.
The problem: which machine produced how much?
Production quantity usually depends on operator-kept records or machine counters; these can be incomplete, delayed or inconsistent. The real performance gap between machines and output per shift is often not clear.
Manual counting takes labour and contains errors; real-time production tracking and machine comparison are not possible.
Counting products off the machine
The camera sees the products coming off the machine (roll, bolt, sheet, pack) and counts them one by one. Each product is marked with its own box and ID (e.g. 'Roll 45', 'Roll 46', 'ID: 12'); the same product is not counted twice. The count appears on the dashboard in real time and is logged with a timestamp.
- Product counting at machine output (roll/bolt/sheet/pack)
- Separate ID per product — no double counting
- Real-time quantity and timestamped log
- Total output per shift and day
Machine and production performance (KPI)
Count data turns into per-machine performance: each machine's hourly/shift output, production speed and downtime are derived. Machines and shifts are compared objectively; which machine/shift is more efficient becomes visible with data.
Production drops and stoppages become visible with timestamps; planning is based on data, not guesswork.
- Hourly/shift output per machine
- Production speed and downtime
- Machine and shift comparison (KPI)
- Early visibility of production drops
Multi-machine / multi-line coverage
Several machines and lines are monitored in one system; each machine's output is counted and compared separately. The plant's total production lands on one dashboard, making balance and bottlenecks between machines visible.
- Separate counting per machine/line
- Plant-wide total production
- Balance and bottlenecks between machines
- Production visibility on one dashboard
On-premise architecture and integration
All analysis runs on a server inside the plant; video does not leave the site. Count and performance data are collected on the dashboard and, if wanted, pushed to an ERP/MES system. When people appear, tracking is anonymous and faces are blurred.
Outcome
Production counting moved from manual, end-of-day records to real-time, automatic and per-machine. How much each machine produced and its performance are visible with data; planning and improvement use measurable data.
- Machine output counted in real time, automatically
- Machine/shift performance logged
- No double counting, tracked by ID
- Production data ready for ERP/MES
Frequently asked questions
How does the system count products?
The camera detects the product coming off the machine (roll, bolt, sheet, pack) and marks and tracks each with its own ID, so the same product is not counted twice. The count appears on the dashboard in real time and is logged.
How does it derive machine performance?
From the count data it calculates per-machine hourly/shift output, production speed and downtime; machines and shifts are compared objectively.
Do I need new cameras or counters?
In most cases no. Existing cameras that see the machine output are enough; no separate counter or sensor is needed on the machine. Field-of-view suitability is assessed in the free pre-assessment.
Is the data pushed to ERP/MES?
Yes, if wanted. Count and performance data are collected on the dashboard and can be pushed to ERP/MES via REST API. Analysis runs on-premise and video does not leave the site.
