/JLS%C2%AE_MasterLogo_Navy.png)
JLS Vision AI — AI vision that turns chaotic product flow into order
Turning chaos into control.
From a random pile to the presentation the next machine needs.
JLS Vision AI is a proprietary, AI-driven machine vision technology from JLS Automation that lets robotic systems pick food products directly from chaotic flow — where items are touching, overlapping, or stacked — without requiring upstream product separation. It identifies which products are accessible and pickable, classifies them, determines their orientation, and tracks them as the pile shifts — directing the JLS robotic system to pick, orient, and organize each item. It processes images at up to 15 frames per second. Developed specifically for JLS robotic systems and debuting at PACK EXPO International 2026.
Why it matters
The chaos won’t change. So the vision had to.
Traditional vision generally needs product separated, spaced, and presented in a clean layer before automation can work with it. Creating that presentation can require vibratory laners, spreader belts, chicaning systems — or two or three people manually sorting and orienting product.
JLS Vision AI changes that preparation step. It sees which product is on top and pickable, classifies its type and orientation, and tells a JLS robotic system how to handle it. The robot then turns the chaotic incoming product into an organized flow for the next machine — even as the pile shifts during picking.
Less manual sorting. Less equipment devoted to separation. A smaller, more efficient path from pile to production-ready flow.
On the actual demo cell
The same packets. Two very different presentations.
Photos from the actual JLS Vision AI demonstration system — the same cell running at PACK EXPO Booth S-2776.
Capabilities
What the vision actually does
JLS Vision AI tells the JLS robotic system how to pick, orient, and place each product so the downstream machine receives it in the presentation it needs.
Detects products that are touching, overlapping, or stacked — and determines which one is on top and can actually be picked, so the robot works straight from bulk flow.
Classifies products by type, determines their orientation, and identifies upside-down items — then tells the JLS robotic system how each item should be handled, placed, or flipped.
As picking disturbs the pile, the vision dynamically reacquires targets and updates the robot in near real time — no stopping to re-image a settled scene.
Supports both 2D and 3D vision, adding height information when an application requires it — stacked product, variable thickness, depth-critical picks.
Excels with products that vary in shape, size, and presentation — tested on steaks, chicken breasts, frozen sausage patties, frozen omelets, cookies, condiment packets, flow-wrapped packages, brick packs, and transparent packaging.
Can be trained on application-specific characteristics — bones or fat content in proteins, surface defects like burn marks on baked goods, tears in packaging. (It complements, not replaces, contaminant-inspection systems like metal detection.)
helping keep vision processing from becoming the system bottleneck The hardware lives inside the JLS control cabinet — low-power, industrial, no external processing rack.
See it live — PACK EXPO International 2026 · Booth S-2776
From a random pile to an organized flow, in front of you
This isn’t a concept or a prerecorded video. JLS Vision AI will be running live at Booth S-2776 — guiding JLS automation as it picks, organizes, and refeeds product in real time. Be among the first to see it in action.
The demonstration is built on a real customer challenge: thin condiment packets that arrive touching and overlapping — a product traditional vision could not reliably resolve. Watch JLS Vision AI direct a JLS robotic system as it:
Identifies the top pickable packet within a random pile
Classifies packets by type and organizes them on the placement conveyor
Determines label orientation — and flips upside-down packets
Reacquires targets as product shifts — while the cell recirculates bypassed product for another pass
The PACK EXPO demonstration cell features a JLS Level 4 washdown design with IP69K-rated components — the same hygienic standard our Level 4 production systems are built to.
The payoff
Simpler lines. Smaller footprints. Fewer bypassed products.
Turning a random pile into organized product flow can otherwise require multiple people, significant conveyor space, and specialized separation equipment. Integrated with a JLS robotic system, JLS Vision AI can perform that work quickly and consistently in a much smaller footprint — preparing product for the next operation without first forcing it into a perfectly separated presentation.
In one proposed customer-system comparison, eliminating upstream separation equipment reduced the estimated line footprint by 45% and projected capital requirements by approximately $1.5 million. Results vary by application.
Depending on the application, additional benefits can include greater throughput, higher robot utilization, fewer bypassed products, and reduced reliance on manual product handling.
Where it runs
Developed specifically for select JLS robotic systems
JLS Vision AI is an application-engineered technology available with select JLS robotic systems. The vision, robot, tooling, controls, product flow, and downstream requirements are engineered together for each application.
It is not sold as a standalone vision package for third-party robots. New-system and existing-system applicability is confirmed through an engineering review and product testing when required.
Prove it on your product
Send us your most chaotic product.
We'll show you what the vision sees.
Bring us the product that defeats conventional vision — overlapping, irregular, transparent, frozen. JLS validates the application on your actual product during the pre-sale process, and every production system is proven at Factory Acceptance Testing before it ships.