Skip to content
RoboFab

Engineering study

Vision GuidedSorting Cell

A camera and a trained model decide where every part belongs

Engineering studyRobotics and AutomationArtificial IntelligenceMechanical Design
Dual arm robot picking items from trays
Reference visual

This is a concept study that shows how RoboFab would engineer this kind of systemIt is not a delivered client project and its images are reference visuals

01The problem

Manual sorting of small parts is slow and inconsistentand the line has no data on what passes through it

02The approach

  1. 01A camera feeds a detection model on an edge computer
  2. 02The model classifies each part and a motion controller drives an arm or diverter to the right bin
  3. 03A dashboard logs throughput and errors for every shift

03Architecture

Systemarchitecture

System architecture

CameraEdge AIMotion controllerArm or diverterDashboard

Technologies

  • Raspberry Pi
  • OpenCV
  • YOLO
  • STM32
  • Stepper and servo control
  • Web dashboard

A delivery would include

  • Mechanical fixture design
  • Dataset and model training
  • Motion control firmware
  • Operator dashboard

04Value

The valueit creates

  • Consistent sorting

    The same decision every time at line speed

  • Measured throughput

    Counts and error rates for every shift

  • Continuous improvement

    Stored images retrain the model over time

Pick and place machine populating a circuit boardReference visual

Need somethinglike this

Start a project

More studies

Otherengineering studies

Other engineering studies