Liquids and solids together
Product floats in wash water, tumbles in flumes and arrives wet on the belt. Reflections, splash and foam confuse simple sensors and 2D cameras.
Food Processing Line Monitoring with LiDARs and AI Vision
Where water, vapor and product meet, count what actually passes.
CRATUS monitors food processing lines with LiDAR and AI vision: item counts, volume, flow and fill levels on belts, flumes, wash tanks, hoppers and kettles. The system is built for the hard cases, where liquids and solids meet, lighting is poor or changing, and the air is full of dust or vapor, and sends results to your line controls and dashboards.
- One system for
- LiDAR
- AI vision
- Edge
- Software
Counts + amountsItems, volume and flow
Built forWater, vapor and dust
- Counts + amountsItems counted and volume measured on the same line
- Wet, vapor, dustBuilt for conditions where simple sensors fail
- LiDAR + AI visionShape from LiDAR, appearance from vision
- Line to dashboardResults to PLC, SCADA and dashboards
The challenge
Why food lines are hard to measure.
Production lines are wet, bright, steamy and fast. The conditions that make food safe to produce make it hard to count.
Lighting you can’t trust
Glare on stainless steel and wet product, shadows, mixed lighting and lines that run outdoors change what a camera sees through the day.
Vapor, steam and dust
Blanchers, cookers and fryers fill the air with vapor; flour, sugar and spices fill it with dust. Lenses fog and foul, and readings drift.
CRATUS combines LiDAR and AI vision, placed and protected for wet, steamy and dusty lines.
See the approachDifficult conditions
How each condition is handled.
Every line is different. These are the conditions we plan for.
| Condition | What goes wrong | How CRATUS handles it |
|---|---|---|
| Product in water | Reflections, splash and floating product | LiDAR and AI vision trained to separate product from water and foam |
| Glare and wet surfaces | Bright spots and missing areas in images | LiDAR unaffected by visible glare; controlled lighting for vision |
| Low or changing light | Counts drift with time of day | LiDAR brings its own light |
| Steam and vapor | Fogged lenses, scattered readings | Placement out of the plume, purge air and filtering |
| Dust (flour, sugar, spices) | Fouled windows, falling accuracy | Purge collars, sealed enclosures and cleaning schedules |
| Wash-down | Water and chemicals on equipment | Sealed, washable enclosures rated for wash-down where required |
| Touching and stacked product | Items merge in 2D | 3D separation with LiDAR plus AI vision |
Sensor selection, placement and protection are set per line during the survey.
What we measure
As diverse as the lines we work on.
Counts, amounts and levels across food and beverage processing.
- Fresh produceTomatoes, potatoes, apples, citrus and more, in flumes, wash tanks and on belts.
- Meat, poultry and seafoodPortions and pieces on belts, trays and chutes.
- Bakery and snacksLoaves, buns, chips and fries; volume and spread on the belt.
- Frozen foodsProduct through freezers and on cold, frosty lines.
- Nuts, grains and seedsVolume and flow of bulk product on conveyors and in bins.
- Liquids and saucesLevels in kettles, tanks and hoppers.
- Packaged goodsPacks, trays and cases counted at the end of the line.
- Ingredients and dry goodsFlour, sugar and powders in hoppers and bins.
Our approach
From the line to the numbers.
LiDAR, AI vision, edge computing and software, engineered as one system by one team.
Sensing
LiDAR measures shape, height and volume in 3D regardless of light and color. AI vision cameras add appearance: size, color and surface. Each is placed and protected for the conditions on your line.
- LiDAR for shape and volume
- AI vision for appearance
- Purge air and wash-down enclosures
Edge computing
An edge computer at the site merges every sensor into one live picture and computes the analytics locally, so results, not raw sensor streams, leave the site over Ethernet, Wi-Fi or cellular.
- Multi-sensor fusion on site
- Analytics computed at the edge
- Ethernet, Wi-Fi or cellular
Software & integration
Counts, rates, volume and levels per line and per shift on a dashboard, sent to your PLC, SCADA or MES over MODBUS, OPC UA or MQTT, with alerts for jams, gaps and low supply.
- Dashboard per line and shift
- MODBUS, OPC UA, MQTT
- Alerts for jams and gaps
AI
AI separates product from water, foam and the belt, splits touching items, and screens size and appearance, trained on your product and your line.
- Product vs water and foam
- Touching items separated
- Size and appearance screening
What we deliver
Food processing line monitoring.
One line or a whole plant.
Item counting
Produce, portions, pieces and packs counted on belts, chutes and in flumes.
Sensing · AI
Volume and flow on belts
Volume and flow rate of bulk product such as fries, chips, nuts and grains.
Sensing · Edge
Flumes and wash tanks
Product load in water, overflow and feed rate into the next step.
Sensing · AI
Hopper, bin and kettle levels
Fill levels of solids and liquids, so supply never runs out unnoticed.
Sensing
Gaps, jams and spacing
Empty belt sections, pile-ups at transfers and uneven spread across the belt.
Software · AI
Size and appearance screening
Size distribution and appearance checks with AI vision.
AI
The CRATUS line monitoring system
Sensors and software for wet, steamy and dusty lines.
Sensor types, placement and protection are chosen per line.
CRATUS
LiDAR sensor
Shape, height and volume
- Unaffected by glare
- No light needed
- 3D separation
Measures what is on the belt or in the tank in 3D, regardless of light and color.
CRATUS
AI vision camera
Appearance
- Size and color
- Surface checks
- Trained per product
Adds what LiDAR cannot see: color, surface and appearance.
ASSET-Rx
Edge computer
ASSET-Rx EDGE
- Sensor fusion
- AI on site
- Line-speed results
Fuses LiDAR and vision and runs the AI models at the line.
CRATUS
Enclosures and purge
For wet and dusty lines
- Wash-down enclosures
- Purge air
- Stainless mounts
Keeps sensors clean and sealed where water, vapor and dust are present.
CRATUS
Line dashboard
Per line and shift
- Counts and rates
- Volume and levels
- Alerts
Line data for production, quality and maintenance teams.
CRATUS
Controls integration
PLC, SCADA, MES
- MODBUS
- OPC UA
- MQTT
Sends results into the controls and systems already running the plant.
One system
Sensors, edge, software and AI from one team
CRATUS surveys the line, selects and protects the sensors, trains the AI on your product and integrates results with your controls.
All CRATUS productsCRATUS line monitoringStack
AI
- Product separation
- Counting
- Screening
Software
- Dashboard
- Alerts
- PLC / SCADA
Edge
- Sensor fusion
- Models on site
- Line-speed
Sensing
- LiDAR
- AI vision
- Enclosures
What to send us
These details let CRATUS plan monitoring for your line.
- Product and what you need: count, volume, flow or level
- Line type: belt, flume, wash tank, hopper, kettle
- Line speed and belt width
- Water, vapor, dust and wash-down conditions
- Lighting, indoor or outdoor
- Mounting options above the line
- Controls the data should reach: PLC, SCADA, MES
- Photos or video of the line
How we work
Surveyed and trained, then installed.
Every line is surveyed for its conditions, and the AI is trained on your product before results go live.
- 01
Line details
Product, measurement needed, line layout, speed and conditions.
- 02
Survey & sample data
Sensor selection and placement; sample data collected from your line.
- 03
Train & install
AI trained on your product; sensors installed with enclosures and purge where needed.
- 04
Commission
Counts and volumes checked against manual counts or weighed batches.
- 05
Operate & support
Live line data, alerts and integration, with retraining as products change.
Standards & practices
Designed for food plants.
Equipment over a food line has to meet hygiene, safety and security expectations.
- FSMA, 21 CFR 117Current good manufacturing practice and preventive controls
- HACCPHazard analysis and critical control points
- 3-A Sanitary StandardsHygienic design of dairy and food equipment
- EHEDGHygienic engineering and design guidelines
- IEC 60529Ingress protection, including wash-down ratings
- NSF/ANSI/3-A 14159Hygiene requirements for processing equipment
- IEC 60825-1Eye-safe laser classes
- ISA/IEC 62443Security for industrial networks and devices
Certification is issued by accredited third parties. CRATUS designs and documents installations to these references and supports your food safety plan; we don’t self-certify.
The basics
What can CRATUS measure on a food line?
Counts, amounts and levels: items on belts and in flumes, volume and flow of bulk product, fill levels in hoppers, bins, tanks and kettles, and gaps, jams and spacing on conveyors.
Why combine LiDAR and AI vision?
Each covers the other’s weak spot. LiDAR measures shape and volume in 3D regardless of light; AI vision sees color and surface. Together they count and measure where either alone would struggle.
Can it count product floating in water?
Yes. LiDAR and AI vision are trained to separate product from water, splash and foam, so product in flumes and wash tanks can be counted and its load estimated.
Conditions
Does steam or vapor stop the system?
No, with the right placement. Sensors are positioned out of the plume where possible, protected with purge air, and readings are filtered for vapor.
What about dust from flour, sugar or spices?
Dusty lines use purge collars, sealed enclosures and a cleaning schedule so sensor windows stay clear.
Can the sensors be washed down?
Yes, where required. Sensors are installed in sealed enclosures rated for wash-down and mounted with stainless hardware.
Integration
Can results go to our PLC or SCADA?
Yes. Counts, rates, volumes and levels can be sent over MODBUS, OPC UA or MQTT to your PLC, SCADA or MES.
How is accuracy checked?
Against your own references. At commissioning, counts and volumes are compared with manual counts or weighed batches, and the AI is retrained as products change.
Working with CRATUS
How do we get a proposal?
Send us the site layout, what you need to measure and how the space is used. A CRATUS engineer replies with questions, a sensor layout and a phased plan.
Can you work with our integrator, security or IT team?
Yes. CRATUS works directly with operations teams, system integrators and IT departments, and provides network requirements, installation instructions and remote support during commissioning.
Talk to a CRATUS engineer
Have a line that is hard to count?
Send us photos or a short video of the line and what you need to measure. A CRATUS engineer will reply with a sensor plan.