Pallet detection engineering guide
Sensor for Pallet Detection: Choose Photoelectric, Ultrasonic, or 3D
There is no universal best pallet sensor. Use a validated optical sensor for a defined conveyor presence point, consider ultrasonic or a wider detection field for recessed plastic pallets and difficult surfaces, and use 3D perception when a robot must locate fork pockets or calculate pallet pose. Start with the decision the machine must make, then prove it on the worst pallet.
- Conveyor presence and leading-edge detection
- Wood, plastic, damaged and wrapped pallets
- ASRS, AGV and AMR pickup verification
Presence or leading edge
Start with through-beam, retro-reflective, a light-band sensor, or a carefully placed background-suppression sensor.
Perforated plastic pallet
Test ultrasonic or a wider optical field when one narrow light spot can pass through recesses and openings.
Distance or stop position
Use a distance-capable sensor and budget the complete reaction and braking travel, not only its catalog range.
Fork pocket and pose
Use a 3D camera or purpose-built pallet perception system when the vehicle needs X, Y, Z, yaw, pitch and roll.
This guide covers industrial pallet detection and positioning. An ordinary photoelectric, ultrasonic or distance sensor is not a personnel-protection device. Safety functions require safety-rated equipment, a validated safety architecture and the applicable machine risk assessment.
Define the output
"Detect the pallet" can mean six different machine decisions.
A sensor can be correct for one decision and inadequate for another. Write the required output before comparing ranges or housing styles.
Presence
Confirm that any approved pallet occupies a conveyor zone, transfer station or rack position.
Leading edge
Create a repeatable trigger when the front of the pallet reaches an indexing or accumulation point.
Stop position
Determine whether the pallet has reached a transfer, lift or handoff position within an allowed window.
Center and overhang
Check width, lateral position or protruding load that could interfere with guides, racks or handling equipment.
Empty, loaded or damaged
Evaluate height or profile. One switching point generally cannot prove pallet condition or load completeness.
Fork pocket and pose
Locate entry pockets and approach angle for a forklift, AGV or AMR using spatial information rather than presence alone.
Technology comparison
Which sensor is best for pallet detection?
The strongest starting point is the technology that can observe a stable pallet feature under the worst geometry and environment. Published range is only one constraint.
| Technology | Good starting use | Why it helps | Main limitations to prove |
|---|---|---|---|
| Through-beam photoelectric | Presence or leading edge across an accessible conveyor | The receiver evaluates a direct beam, so detection is less dependent on pallet surface reflectivity. | Two-sided wiring and alignment; a narrow beam may pass through pallet openings unless the sensing height is chosen correctly. |
| Retro-reflective or light-band | One-sided wiring, irregular leading edges, open or damaged pallet profiles | A reflector closes the optical path. A wider light band can reduce the risk that one point passes through a hole. | Reflector space, alignment, film reflections, contamination and the exact minimum interruption required. |
| Diffuse / background suppression | One-sided presence at a defined foot, block or deck position | Simple installation; BGS can help reject a conveyor or structure behind the target by distance. | Target color, gloss, angle and irregularity still matter. The selected target feature must remain inside the sensing window. |
| Ultrasonic | Recessed plastic pallets, difficult colors, or applications where optical contrast is unreliable | Uses an acoustic echo rather than reflected brightness. A broader sound field may bridge small openings. | Blind zone, beam width, response time, angled surfaces, nearby structures, air movement and the minimum spacing between objects. |
| Laser distance / time-of-flight | Distance window, approach measurement or position confirmation | Provides a measured distance or a tightly defined switching point. | Spot footprint, pallet gaps, dark or reflective surfaces, background selection, update rate and total control delay. |
| Light grid, array or 2D LiDAR | Irregular leading edges, overhang, profile or a larger monitored field | Observes more than one point, reducing dependence on a single slat, block or opening. | Resolution, field height, scan/update time, occlusion and how the controller interprets the field data. |
| 3D camera / pallet perception | Fork-pocket location, six-degree pose, damaged or misaligned pallet pickup | Provides spatial data for X, Y, Z and orientation instead of one binary signal. | Field of view, minimum/maximum working distance, scene occlusion, algorithm coverage, integration and validation dataset. |
| Inductive proximity | Metal pallet, steel carrier or an intentional metal flag | Does not depend on optical appearance and is useful at close range. | It does not directly detect ordinary wood or plastic. Target metal, size, set distance and surrounding metal affect the result. |
OMRON's photoelectric selection guidance asks engineers to evaluate target size, shape, color, material, surface, velocity, background and environment, not only sensing distance.[1]
Where XSZ sensors fit
Use a switching sensor when the machine needs one clearly defined event.
XSZ photoelectric sensors can support pallet presence, leading-edge and position checks when the target feature and mounting geometry are controlled. Background suppression can be useful when a pallet block is in front of a nearby conveyor or machine surface. Through-beam and retro-reflective arrangements can provide a stronger optical path when both sides, or a reflector, are available.
Do not ask one switching sensor to infer pallet pose, pocket location, damage, overhang and load condition simultaneously. Those jobs require a larger sensing field, multiple coordinated points, a distance profile or 2D/3D perception.
- Confirm the closest and farthest pallet feature, not only the conveyor width.
- Test the darkest, brightest, most damaged and most misaligned approved pallet.
- Match voltage, PNP/NPN, NO/NC, connector and PLC input behavior.
Target variation
Choose around pallet construction, not the material name alone.
"Wood pallet" or "plastic pallet" does not define a stable optical or acoustic target. Deck openings, blocks, stringers, damage, film and transport orientation decide what actually crosses the sensing field.
Natural variation creates gaps and weak points.
Knotholes, broken boards, missing blocks, oil, labels and film can change the signal. Aim at a feature present on every approved pallet, or widen the monitored field.
Color is only one challenge.
Molded ribs, large recesses, glossy surfaces and open decks can make a point sensor alternate between target and no target. Test beam width, sound field and transport direction.
A wider acoustic field can bridge recesses.
Pepperl+Fuchs documents ultrasonic sensing as one route for special plastic pallets with many recesses and difficult optical conditions.[2] Confirm blind zone and timing on the exact model.
Mounting geometry
Put the sensing field where every approved pallet creates one stable event.
The most reliable installation begins with a physical drawing: pallet feature, beam or sound field, conveyor structure, travel direction, height tolerance and background.
Build the detection plane in four layers.
Select a block, stringer, deck edge or loaded profile that exists on every allowed pallet and remains visible through the full lateral and vertical tolerance.
Compare the effective light spot, light band, sound cone or scan field with holes, slat spacing and the smallest continuous target feature.
Document what the sensor sees with no pallet: roller, chain, floor, rack, reflector, crossbar or open aisle. A stable "clear" state is as important as target detection.
Protect the active face and cable from impact, but preserve access for alignment, cleaning, teach-in and replacement without changing the calibrated position.
Application starting point
Choose the condition most likely to cause a missed pallet.
This selector identifies a technology family to investigate first. It is not a model guarantee; final selection requires drawings, pallet samples, environmental limits and machine testing.
Select the hardest condition
Start with a beam-interruption or wider optical field
For mixed wooden pallets, a through-beam or retro-reflective arrangement avoids depending on one diffuse return. A wider light band or carefully chosen target height helps prevent openings, damage and irregular edges from splitting the signal.
Watch: A narrow point beam may pass through a pallet opening or switch twice across a broken edge. The geometry must be proven along the complete pallet length.
Review the relevant sensor familyLong-distance selection
A range number does not prove reliable pallet detection.
At longer distances, the installation becomes an optical or acoustic geometry problem. Check the complete return path, field footprint and environmental margin using the exact device documentation.
Confirm the usable margin, not the maximum line in a table.
A sensor's published range is measured under stated target, reflector and environmental conditions. The real installation can lose margin because the target is smaller, darker, angled, perforated, wet or partially hidden.
For a point-return sensor, measure the beam footprint at the pallet and confirm that it stays on a continuous feature throughout skew and lift variation. For retro-reflective or through-beam sensing, confirm reflector/receiver size, alignment tolerance and the smallest interruption caused by every pallet type.
When the job is distance measurement rather than presence, check repeatability, linearity, update rate and the controller's usable distance window. A precise distance value is still wrong if the spot is reading through an opening to the rack behind it.
Field footprint
Determine spot, band, sound cone or scan-field size at the actual distance and compare it with pallet openings.
Target return
Test the weakest color, surface, angle and smallest continuous feature, not a clean white reference target.
Background
Identify floors, racks, conveyor parts, reflectors and loads that can enter the field when the pallet is absent.
Operating reserve
Validate after realistic contamination, temperature change, vibration and alignment tolerance, not only immediately after teach-in.
Harsh environment
Dust, moisture, cold and film do not have one universal sensor answer.
Ingress rating, sensing principle and mounting design solve different problems. Match each environmental threat to a test and a maintenance response.
Dust and lens fouling
Optical reserve falls as the active face or reflector becomes contaminated. Use a protected orientation, accessible cleaning point and contamination diagnostics when available. Ultrasonic may be a candidate for difficult optical conditions, but heavy buildup on its face and nearby structures still require testing.
Water and washdown
Verify the complete installed assembly: housing, connector, cable gland, bracket, cleaning chemical and jet direction. A high ingress code does not automatically prove chemical compatibility or a stable optical path through droplets.
Cold, condensation and frost
Check the specified operating temperature and repeated transitions through the dew point. A sealed housing can still have an obscured lens. Consider supported heated optics, air purge, sheltered mounting or maintenance cycles where required.
Black or reflective wrap
Diffuse optical return can change sharply with dark film, glare and angle. Test polarized or application-specific optical modes, beam interruption, ultrasonic or a larger field against the actual wrap, seams and load overhang.
Shock and vibration
Protect the sensor from pallet, fork and debris impact. Use a rigid repeatable bracket, strain relief and a replacement datum so service work does not move the calibrated field.
Ambient light and interference
Test normal LED lighting, sunlight, adjacent optical sensors and nearby ultrasonic devices. Use supported anti-interference features, physical separation or staggered operation where required.
Position planning
Estimate how far the pallet moves after the sensor detects it.
Detection delay, input and logic delay, communication delay and mechanical braking all consume distance. This planning estimator compares the entered total with the available distance to the desired stop point.
Stop-position margin estimator
The entered values leave 60 mm before the desired stop point.
The pallet travels about 20 mm during the entered sensor and control delay, then a further 120 mm during the entered mechanical stop.
Planning aid only. Enter measured worst-case braking travel and verified total delay. The calculation does not include load-dependent braking variation, conveyor slip, safety factors, control architecture or legal/safety requirements.
AGV and AMR pickup
A binary presence sensor cannot calculate a pallet's approach pose.
A fork-tip or carriage sensor can confirm that a pallet or load is present at a defined point. It cannot by itself tell a vehicle how far left, right, high or rotated the pallet is. For autonomous pickup, the perception system may need the pallet's X, Y and Z position plus yaw, pitch and roll.
ifm describes a 3D pallet detection workflow that captures point-cloud data, calculates six-degree pose and supplies left/right pocket locations to the vehicle.[5] This is a different class of task from switching a conveyor zone.
Electrical integration
Carry a clean pallet event from the target to the machine decision.
A stable sensor LED is only the first layer. Confirm the electrical output, controller input, timing logic and fault behavior as one signal chain.
Sensor output
Match voltage, PNP/NPN, NO/NC or push-pull behavior, maximum load and connector pinout to the input architecture.
Input capture
Verify filter, scan, update or pulse-capture behavior so the shortest valid pallet event reaches the program.
Machine logic
Define which edge establishes presence, how the zone clears and what happens if a pallet stops on the transition.
Diagnostics
Use contamination, signal-strength or teach alarms when supported, and distinguish a blocked path from wiring or device fault.
Replacement
Record model, settings and bracket datum. IO-Link parameter restore requires compatible device support and correctly configured master data storage.[6]
For wiring fundamentals, review NPN vs PNP sensor outputs and NO vs NC sensor behavior before selecting the replacement configuration.
Commissioning and acceptance
Validate the complete pallet population, not one demonstration sample.
Acceptance should prove the sensor, mount, controller and machine action under the worst approved production conditions.
Build the pallet library
List every pallet design, material, color, entry orientation, empty/loaded state and wrap condition that the machine must accept.
Include real damage
Test chipped blocks, broken boards, warped decks, oil, labels, hanging film and the allowed maintenance limit.
Challenge geometry
Run minimum and maximum height, lateral offset, skew, bounce, overhang and both transport orientations where applicable.
Challenge the environment
Test realistic dust, droplets, cleaning state, temperature transitions, lighting, vibration and adjacent sensors.
Observe every signal layer
Compare sensor status, raw electrical output, PLC input, program state and final conveyor or vehicle action.
Run maximum dynamics
Use maximum approved speed and load, minimum spacing and measured stop behavior rather than a slow manual pass.
Test abnormal flow
Stop on the sensing edge, restart, reverse if allowed, create a jam, remove a pallet and test power-cycle behavior.
Freeze the baseline
Record model, firmware, settings, mounting dimensions, input filter, logic, sample set and accepted results.
Troubleshooting
Use the failure pattern to identify the missing detection margin.
Change one variable at a time and observe both the sensor output and the PLC input. Avoid hiding an unstable target signal with an arbitrary timer.
| Symptom | Likely cause | Check first | Corrective direction |
|---|---|---|---|
| Wood pallet is detected intermittently | Point beam crosses a slat gap, knothole, missing block or strong surface variation. | Mark the beam path on each pallet design and observe output along the full travel. | Move to a continuous feature, change height/orientation, or use a wider band or field. |
| Plastic pallet flickers | Glossy ribs and recesses alternate between strong return and no target. | Compare signal at every molded feature, color and direction. | Evaluate ultrasonic, beam interruption, a larger field or application-specific optical mode. |
| False pallet on an empty conveyor | Background, roller, rack, reflector, hanging film or contamination enters the sensing window. | Inspect the clear-state signal at all conveyor and lighting conditions. | Reposition, angle, shield, narrow the distance window or change sensing principle. |
| Sensor switches, PLC does not | Output/input mismatch, wiring issue, filter delay, scan behavior or pulse shorter than capture capability. | Measure the electrical output and controller input simultaneously. | Correct PNP/NPN and common wiring, configure the input, or use a supported faster capture path. |
| Pallet stops too late | Sensing point is too close, total control delay is underestimated, or braking varies with load. | Measure delay and stop travel under worst load and speed. | Move the sensing point upstream, improve reaction path, or redesign the position-control method. |
| Works clean, fails after a shift | Lens/reflector fouling, bracket movement, condensation or cable damage reduces reserve. | Compare current signal/position with the commissioned baseline. | Improve protection and access, add diagnostics, repair the mount/cable and define maintenance intervals. |
| AMR sees pallet but misses pockets | Presence is confused with pose; pockets are occluded, wrapped, damaged or outside trained coverage. | Review point cloud, confidence, field of view and exception logs. | Use validated pallet-pose perception and define unsupported pallet/scene handling. |
For a broader diagnostic workflow, use the XSZ Sensor Troubleshooting Guide.
Prepare the requirement
Send these facts for a useful pallet sensor recommendation.
A model number can be selected faster when the application is described as measurable limits rather than "long distance" or "harsh environment."
Detection decision
Presence, leading edge, stop position, overhang, empty/loaded, damage or fork-pocket pose.
Pallet population
Material, dimensions, blocks/stringers, openings, colors, damage, wrap and transport orientations.
Distance and tolerance
Closest/farthest target, mounting space, lateral/height movement, skew and desired switching window.
Conveyor dynamics
Maximum speed, pallet spacing, load range, stop distance and the timing requirement at the controller.
Environment
Temperature cycle, dust, water, oil, chemicals, washdown, frost, vibration, impact and ambient light.
Electrical interface
Supply voltage, PNP/NPN, NO/NC, analog/IO-Link need, connector, cable and PLC/robot input type.
Turn the pallet drawing and machine limits into a testable sensor choice.
Send XSZ the pallet photos or drawings, sensing distance, mounting view, line speed, environment, voltage, output type and required quantity. We can help narrow the optical sensing mode and sample configuration for validation.
Related XSZ resources
Continue from pallet geometry to the exact sensing and wiring decision.
Frequently asked questions
Sensor for pallet detection FAQ
What is the best sensor for pallet detection?
For a defined pallet-presence or leading-edge point on a conveyor, start by evaluating through-beam, retro-reflective, wider light-band or background-suppression photoelectric sensing. Perforated plastic pallets or difficult optical surfaces may justify ultrasonic testing. Distance windows need a distance-capable sensor, while fork-pocket location and pallet pose require 3D perception. The best model is the one validated against every approved pallet and the worst environment.
Can a photoelectric sensor detect wooden pallets reliably?
Yes, when the sensing field crosses a feature that remains present throughout the allowed pallet population and movement tolerance. Wood can include slat gaps, knotholes, broken boards, oil, labels and film. A point beam may pass through an opening, so compare its footprint with the pallet geometry and consider beam interruption or a wider optical field for irregular pallets.
Should I use photoelectric or ultrasonic sensing for plastic pallets?
Test both when molded recesses, glossy surfaces or changing colors make a point optical return unstable. Ultrasonic sensing does not depend on reflected brightness and a wider sound field may bridge openings, but blind zone, response time, beam width, angled surfaces and nearby structures still matter. A through-beam, retro-reflective or wider optical field may also solve the application.
How do I detect a pallet with shrink wrap or reflective film?
First decide whether the sensor should detect the pallet structure, the complete wrapped load or both. Test polarized or application-specific retro-reflective optics, through-beam interruption, background suppression, ultrasonic or a larger sensing field against the real film, seams, overlap, tension, load overhang and approach angle. Do not assume one clear or glossy sample represents every wrapped load.
Can one sensor distinguish an empty pallet from a loaded pallet?
A single point can distinguish empty from loaded only when a defined load feature always crosses that point and the pallet itself never does. Mixed loads, overhang and variable height usually require an additional height point, light grid, distance profile, 2D LiDAR or 3D camera. Keep pallet presence and load-profile decisions separate in the control logic.
What sensor is needed for AGV or AMR pallet pickup?
A switching sensor can verify presence or final pickup at a defined point, but autonomous approach normally needs spatial perception. A purpose-built 3D pallet detection system can calculate pallet position and orientation and locate fork pockets. Validate field of view, working distance, pallet designs, damage, wrap, occlusion, lighting and the vehicle's exception behavior.
How far before the stop point should a pallet sensor be mounted?
Determine it from the measured worst-case system behavior, not a universal distance. Add pallet travel during sensor, input, logic and communication delay to the measured mechanical stopping travel, then include the machine's required tolerance and engineering margin. Repeat the measurement at maximum speed and worst load, because braking and slip can change with operating conditions.
Does IP67 or IP69K make a sensor suitable for cold storage?
Not by itself. An ingress rating addresses specified dust and water exposure; it does not prove the operating temperature, resistance to thermal cycling, prevention of condensation/frost on the sensing face, chemical compatibility or stable detection through droplets. Check the complete model specification and connector, then test repeated transitions through the real temperature and dew-point conditions.
Evidence and media
Technical references and image credits
- OMRON Industrial Automation, Overview of Photoelectric Sensors - selection factors for through-beam, retro-reflective, diffuse and background-suppression sensing.
- Pepperl+Fuchs, Pallet Detection on Roller Conveyor Systems: Optical or Ultrasonic? - application comparison for standard wooden and recessed plastic pallets.
- Leuze, Optimized for Irregular Objects, Pallets and Polybags - wider light-band use for irregular objects and pallets with openings.
- SICK, Leading Edge Detection on Conveying Equipment - target, speed, vibration and larger-field detection considerations for pallets and irregular loads.
- ifm, Pallet Detection System Pick - 3D point-cloud processing, six-degree pose and fork-pocket location for mobile robots.
- IO-Link Community, Interface and System Specification V1.1.4 - master/device data storage and parameter restore mechanism.
- ifm, Ultrasonic Sensor Technology and Installation Guidelines - blind zone and installation considerations.
Image credits
- Pallet detector hero, optical/ultrasonic comparison and plastic-pallet application: Pepperl+Fuchs.
- XSZ background-suppression sensor: XSZ Sensor.
- Wooden pallet stacks: Timothy Huliselan / Pexels.
- Blue plastic pallets: Mathias Reding / Pexels.
- 3D pallet perception: ifm.