Photoneo AnyPick Module for Boxes (Depalletization)

The AnyPick module for Boxes (Depalletization) is Photoneo's AI-powered depalletization engine. Photoneo states it localizes and depalletizes over 500,000 types of boxes at a speed of 1300 units per hour, with up to 99.7 percent pick rate. It handles randomly palletized boxes with no CAD model and no fixed stacking pattern.

Published Performance

Note that some third-party listings still circulate older performance figures from Photoneo's earlier Universal Depalletizer material. The values above are Photoneo's current published figures for this module.

The Problem It Solves

Randomly palletized boxes have historically been the hard case in logistics automation. A conventional vision system needs to know what it is looking for. A mixed pallet arriving with unfamiliar cartons, inconsistent stacking, and no manifest gives it nothing to match against.

One specific failure mode is worth naming, because Photoneo calls it out directly. Tightly packed boxes are difficult for automation systems because the system struggles to tell the line separating two adjacent boxes from a line contouring the opening or seam of a single box. Get that wrong and the robot either tries to pick two boxes as one or fails to find a pick at all. Photoneo states its algorithms handle this case.

Photoneo also notes that the AI learning engine is not limited to boxes. The robust AI-learning engine allows picked goods to be expanded to bags, bottles, and beyond.

The AnyPick Family: Which Module You Need

Photoneo offers three AnyPick modules. They are not tiers. They are trained for different goods, and the right one is determined by what actually comes off your trucks.

Depalletization or Delayering

These are two different processes and Photoneo sells a separate module for each.

If your downstream process needs individual cartons singulated, depalletization is the module. If you are moving whole layers to another pallet or conveyor and the layer stays together, the LayerPick delayering module is the faster path. Some operations run both.

Integrator Features

• Gripping point computation. Photoneo's AI post-processing includes a dedicated approach for computing gripping points on localized boxes, rather than returning only a position.
• Configurable object prioritization. Photoneo notes this ensures stable sorting while still using results computed in the previous run, and calls out depalletization specifically as a time-critical application where consistent sorting is required. In practice this is what stops the robot re-deciding pick order between scans.
• Automated diagnostics. The system can automatically save 3D scans captured during instances of suboptimal performance. Photoneo describes this as opt-in targeted data collection that gives integrators the diagnostic material to analyze and adjust parameters, which is most valuable during initial configuration.
• Vision-only or vision-guided. Photoneo notes the modular nature of the depalletization tool means it can be used as a standalone vision solution, or upgraded with path planning for full vision-guided robotics.

What You Need Alongside It

• Host platform, required. Bin Picking Studio or Locator Studio. AnyPick runs inside one of these.
• 3D sensor, required. For full-pallet coverage, the PhoXi 3D Scanner Gen3 in XL covers 2008 x 1656 mm at its 2331 mm sweet spot, which brackets a standard 48 by 40 inch pallet with margin. Smaller footprints may suit the L.
• Sensor compatibility. Photoneo states the AnyPick module for Boxes is supported and trained in combination with PhoXi 3D Scanners. Confirm sensor selection with us before ordering if you are considering a MotionCam-3D for this application.
• Calibration hardware. A marker board or calibration ball for the sensor-to-robot transform.
• Companion modules. The LayerPick delayering module if you also move whole layers. AnyPick for Totes or AnyPick Custom if your goods extend beyond cartons.
• Service coverage. Repairs-only and advance replacement plans for the sensor and vision controller.

Specifying AnyPick for Boxes: Automation Distribution Application Notes

The information above is drawn from Photoneo's published documentation. This section is Automation Distribution's guidance, based on that documentation and our experience specifying these systems. It is not manufacturer specification.

• 1300 units per hour is a vision figure, not a cell figure. Localization speed is one term in the cycle. Robot reach and speed, gripper cycle, place position, and conveyor handoff all sit downstream of it. Treat the published throughput as the ceiling the vision system will not be the reason you miss, then size the rest of the cell against your actual target rate.
• Audit what actually arrives before choosing a module. Sites that describe their inbound as "boxes" often turn out to be running a meaningful share of totes, shrink-wrapped bundles, or polybags. Because Photoneo supports switching between networks on the fly, the answer is often more than one module rather than the wrong single one. Worth a week of counting before you commit.
• Gripper selection is the other half of the project. The module returns computed gripping points on localized boxes. Whether those points are reachable and holdable is a vacuum or mechanical gripper question, and it depends on carton top surface, weight, and whether cases are shrink-wrapped or damp. Specify vision and end effector together.
• Budget for retraining on unusual inventory. The pre-trained network covers a very large box population, but operations running non-standard packaging see the value of custom training. Photoneo offers retraining tiers sized by scan count. Ask us about these at specification time rather than after go-live.
• Turn on automated diagnostics from day one. The opt-in scan saving costs nothing during commissioning and gives you the failure cases you will otherwise be trying to reproduce from memory three weeks later.

Frequently Asked Questions

What is the Photoneo AnyPick module for Boxes?

It is Photoneo's AI-powered depalletization software module. Photoneo states it localizes and depalletizes over 500,000 types of boxes at 1300 units per hour with up to 99.7 percent pick rate. It runs as a module inside Bin Picking Studio or Locator Studio and requires no CAD model of the boxes.

Can I buy the AnyPick module on its own?

No. AnyPick is a software module that extends Bin Picking Studio and Locator Studio. A working deployment needs one of those platforms plus a Photoneo 3D sensor. Contact Automation Distribution and we will confirm the full configuration.

Do I need a CAD model of the boxes?

No. AnyPick uses AI localization with a neural network pre-trained on thousands of boxes, rather than CAD-based matching. It also does not require boxes to be stacked in a known pattern.

What is the difference between depalletization and delayering?

Depalletization is robotic unloading of individual boxes one at a time. Delayering is robotic unloading of entire layers of goods while keeping the items within each layer intact. Photoneo supplies AnyPick for Boxes for the first and the LayerPick module for the second.

Can it handle totes as well as boxes?

Totes are covered by a separate module, AnyPick for Totes, which Photoneo built specifically for standardized totes and which localizes them whether empty, filled, or with rims bent from extensive use. Because Photoneo's AI engine supports multiple neural networks in one solution with on-the-fly switching, a single cell can run both.

What if my goods are not boxes or totes?

Photoneo offers AnyPick Custom for applications involving a wide variety of mixed goods, which includes custom training of neural networks to match the application. Photoneo also notes the AI learning engine allows picked goods to be expanded to bags, bottles, and beyond.

Can it pick tightly packed boxes?

Photoneo states its algorithms recognize boxes that are very tightly packed, and identifies why this case is hard: an automation system has to distinguish the line separating two adjacent boxes from a line contouring the opening of a single box.

Does it require a robot?

Not necessarily. Photoneo notes the modular nature of the depalletization tool means it can be used as a standalone vision solution, or upgraded with path planning for use in vision-guided robotics.