PROCESSING FLOW
Robot Vision Processing Flow — Trigger to Pick Verification
From capture trigger to robot pick — a practical breakdown of the AimFactoryCore vision and robot operation cycle as operators see it on the factory floor.
OPERATION
How vision operates inside AimFactoryCore
AimFactoryCore bundles setup, operation, and verification in one continuous cycle.
Camera & Lighting Setup
Exposure, gain, white balance, and lighting timing are configured per model inside the Recipe. No manual adjustments on model change.
ROI Definition
Vision area, exclusion zones, and pick target regions are defined per model. Only the relevant area is processed each cycle.
Pattern Learning
Target part appearance is trained with sample images. Rotation, scale, and partial occlusion tolerance are built in.
Pose Estimation
Position (X/Y) and rotation angle (θ) are computed per candidate. Pick feasibility ranking is applied automatically.
Coordinate Calibration
Image coordinates are converted to robot coordinate space using 9-pose calibration. Residual verification ensures ±0.05 mm accuracy.
Result Output
Pick coordinates, trigger signals, and status flags are delivered to the robot and PLC via TCP/IP or IO in one standard format.
SCENARIO
What happens in one cycle
From one trigger in to pick verification complete — 5 stages as seen from the factory floor.
- 1
TRIGGER
The line sends a "capture now" signal to vision
When the PLC or robot controller issues a capture trigger, AimFactoryCore vision captures a single frame.
- 2
RECOGNITION
Vision locates the part and calculates its pose
Within the ROI, position, rotation angle, and pick candidates are computed in one cycle.
- 3
TRANSFORM
Image coordinates are converted to robot coordinates
Calibration data converts X/Y/θ coordinates to mm units the robot accepts directly.
- 4
DELIVERY
AimFactoryCore sends coordinates to the robot
Whether TCP/IP or IO, coordinates, triggers, and sequences are organized via a standard interface and delivered.
- 5
VERIFICATION
Vision re-checks the pick result
Vision verifies pick feasibility and actual pickup status, then either advances to the next cycle or triggers recovery.
DEMO
Processing flow in action
Vision recognition through to pick — actual operation footage.
AIVE Multi-Model Precision Test
A full cycle: vision recognizes parts, converts to robot coordinates, and follows through to Pick & Place.
START
Let's walk through the flow with your parts.
Share your part samples and current robot — we'll review the recognition, coordinate conversion, and pick flow together.