Measurement goal
Height/volume, 6D pose, collision-aware pick, or mere presence?
Knowledge · Architecture
Select a 3D camera from the task, not the brochure.
Height, volume, pose or messy bins need different depth data than a pretty point-cloud demo. Start from what the robot or measurement actually needs.
Questions first
What must 3D deliver?
Without answers on tolerance, working distance, surface and takt you pick technology from marketing, not feasibility.
Scene
Matte/reflective, open/bin, fixed presentation or chaos?
Environment
WD, FOV, ambient light, dust, cycle time and PLC/robot interface.
Selection path
From requirement to technology class
Then compare stereo, structured light/profilers and RGB-D, see the sensing guide. Product rankings without your scene are not usable.
-
01
Rule 2D in or out
If height/orientation is not required, 2D often stays faster and cheaper.
-
02
Accuracy vs coverage
Higher depth precision often costs FOV, speed or price, make the trade-off explicit.
-
03
Test the surface
Dark, shiny or transparent materials break many ‘standard’ 3D setups.
-
04
Integration path
Include SDK, robot controller, calibration (hand-eye) and logging before purchase.
Samples before shortlist
Share part, bin/presentation and tolerance, we say which 3D class is worth testing.