Robot / Cobot

• Robot guidance using a vision system involves detecting the position and orientation of parts or features in real-time to guide robotic arms for precise pick-and-place, assembly, welding, or handling tasks.
• The system translates visual data into robot coordinates, allowing robots to handle parts with flexibility, even when part positions are not fixed.

Types of AI-Based Object Detection

2D Vision Guidance: Uses 2D camera data to locate objects in X-Y and orientation (θ)

2.5D Vision Guidance: Adds height or surface profile estimation

3D Vision Guidance: Uses stereo vision or depth cameras for full X-Y-Z + rotation

Visual Servoing: Real-time feedback from vision to adjust robot motion

Fiducial/Marker Detection: Uses visual marks to determine robot reference position

Typical Applications

• Part pickup from conveyor or bins (2D/3D guidance)
• Bolt or hole alignment for automatic assembly
• Flexible robotic loading/unloading of components
• Robotic path correction in dynamic environments
• Seam tracking in welding or adhesive application

Solution Provided

• Object localization and pose estimation
• Calibration between the camera and robot coordinate systems
• Generation of robot-compatible position data (e.g., TCP, offset)
• Integration via TCP/IP, Modbus, or robot-specific APIs
• Compensation for part shift, orientation, or height variation

• Robot pick and place using vision enables flexible object pickup from variable positions or orientations and placing them accurately at the target location.
• Vision assists the robot in locating parts and adjusting grip coordinates based on the current scene.

Types of Inspection

Fixed Camera System: Camera mounted above with fixed FOV

Hand-Eye Camera System: Camera mounted on robot for close-range pickup

2D Pick and Place: Guidance using 2D image coordinates

3D Pick and Place: Bin picking using depth or stereo cameras

Typical Applications

• Random part picking from a bin or a conveyor
• Component feeding for assembly lines
• Depalletizing and tray unloading
• Position correction before pick
• Dynamic stacking/unstacking tasks

Solution Provided

• Vision-guided pick coordinates generation
• Part identification and position adjustment
• Grip point selection and orientation correction
• Robot program synchronization with vision
• Support for multiple robot brands and interfaces

• Robot-based inspection systems use robotic arms equipped with vision sensors to move around complex parts and inspect various features dynamically.
• This allows for flexible, multi-angle inspection of components that are large, irregular, or positioned variably.

Types of Inspection

Fixed Camera, Moving Part: Part moved into camera view for inspection

Robot-Mounted Camera: Camera moves with the robot for flexible view angles

Synchronous Inspection: Robot movement synchronized with the inspection cycle

3D Inspection: Uses structured light or laser scanners for profile checking

Typical Applications

• Multi-view surface defect inspection
• 3D scanning of complex parts
• Weld seam inspection using a robot-mounted camera
• Inspecting internal cavities or profiles
• Automated quality inspection in flexible production lines

Solution Provided

• Camera-robot tool frame calibration
• Path planning for coverage and visibility
• Integration with PLC or SCADA for inspection logging
• Support for multiple modalities (2D, 3D, thermal)
• Automated decision making (OK/NG) with inspection feedback

• Vision-based palletization uses image processing to identify product locations and guide robots for systematic stacking or unloading on pallets.
• The system adapts to dynamic product positions, orientations, and stack configurations in real time.

Typical Applications

• Box or tray stacking on pallets
• Depalletizing mixed products from the stack
• Flexible pattern-based layer stacking
• Height or tilt correction during stacking
• Vision-guided placement with orientation alignment

Types of Inspection

2D Vision Palletization: Basic location and orientation detection in 2D

3D Vision Palletization: Height mapping and stacking coordination

Pattern Recognition: Supports various stacking patterns

Dynamic Adjustment: Real-time adjustment for misaligned or shifted items

Techniques

• Real-time part localization before stacking
• Adaptive stacking strategy with height and alignment logic
• Support for random or structured loading
• Robot path planning based on vision feedback
• Automatic layer separation and pattern change