How can I use vision inspection software to work with PLCs and robots in an existing line?

You can integrate vision inspection software with PLCs and robots in an existing line by treating the vision system as a “decision engine” that reads images, makes pass/fail or measurement decisions, and then exchanges those results with the controls layer and robot controller in real time.

Typical integration architecture

1. Camera + lighting + lens
   Captures the part consistently.

2. Vision inspection software
   Performs:

   • presence/absence checks
   • dimensional measurements
   • alignment/orientation detection
   • OCR/OCV
   • defect detection
   • color or surface inspection

3. PLC
   Handles line logic:

   • part present
   • trigger signal
   • reject gate timing
   • interlocks
   • machine state
   • alarms

4. Robot controller
   Uses vision outputs for:

   • part pick location
   • bin picking
   • guidance/offset correction
   • target selection
   • inspection path adjustment

Common ways the vision software works with PLCs

Vision software usually communicates with PLCs through industrial protocols such as:

• EtherNet/IP
• PROFINET
• Modbus TCP
• EtherCAT
• OPC UA
• Discrete I/O for simpler systems

What gets exchanged

From PLC to vision:

• trigger to inspect
• part recipe ID
• product type
• conveyor stop/slow signal
• job selection

From vision to PLC:

• pass/fail result
• measurement values
• part location/offset
• defect codes
• confidence score
• image captured / inspection complete
• retry / reject / alarm signals

Common ways the vision software works with robots

For robots, the vision system can send:

• X/Y/Z position offsets
• rotation angle
• pick pose
• multiple candidate locations
• part orientation
• quality or classification results

The robot may receive data via:

• robot-native Ethernet interfaces
• socket communication
• OPC UA
• vendor-specific APIs
• direct integration plugins from the vision software

How to retrofit it into an existing line

1. Define the inspection task

Be clear on what the vision system must do:

• detect a part
• verify assembly
• measure a feature
• locate a part for a robot
• sort good/bad parts

2. Identify existing PLC and robot interfaces

Check:

• PLC brand/model
• available comms protocols
• I/O points
• scan time
• available tags/registers
• robot controller brand and supported messaging

3. Place the camera station

Choose a location with:

• stable part presentation
• controlled lighting
• minimal vibration
• enough working distance
• repeatable trigger point

4. Decide trigger and timing method

The PLC typically:

• senses part arrival
• triggers the vision camera
• waits for inspection done
• then acts on the result

This can be done with:

• photoeye trigger
• encoder-based trigger on conveyors
• PLC handshake bits
• software trigger from the PLC

5. Configure handshake signals

A simple PLC-to-vision handshake might look like:

• PLC sets Inspect_Req = 1
• Vision software reads the part
• Vision sets Busy = 1
• Vision completes inspection
• Vision sets Result_Ready = 1, Pass = 1/0
• PLC reads result and resets request

For robots, a similar handshake may include:

• New_Pose_Ready
• Pose_Valid
• Robot_Ack
• Job_Complete

6. Map inspection results to control actions

Examples:

• Pass: continue conveyor
• Fail: activate reject pusher
• Misaligned part: stop line and alarm
• Robot pick offset: adjust pick coordinates
• Wrong part type: divert to rework lane

7. Test cycle time

Make sure the complete loop fits the line speed:

• image capture
• processing time
• PLC communication latency
• robot motion planning
• reject timing

Best practices

• Use good lighting first; software cannot fully fix poor imaging.
• Keep PLC logic simple: let the vision system decide, PLC handle machine actions.
• Add timeout handling for missed triggers or failed inspections.
• Use recipe management if multiple SKUs run on the line.
• Log images and results for troubleshooting and traceability.
• Validate with real production parts, not just samples.
• If using robots, calibrate the camera-to-robot coordinate system carefully.

Example workflow

1. Sensor detects a part on the conveyor.
2. PLC sends a trigger to the vision system.
3. Vision captures an image and finds the part’s position and condition.
4. Vision sends:
   • Pass/Fail
   • defect code
   • offset coordinates
5. PLC uses pass/fail to run reject logic.
6. Robot uses offset coordinates to pick the part accurately.

If you’re choosing software

Look for vision software that supports:

• PLC protocol drivers
• robot communication libraries
• configurable I/O mapping
• recipe handling
• image logging
• barcode/OCR tools
• calibration tools
• offline simulation/debugging

If you want, I can also give you:

1. a sample PLC handshake diagram,
2. a robot vision pick-and-place example, or
3. a retrofit checklist for an existing production line.