VISION DIFFERENCE
Machine vision vs robot vision — the two are not the same.
Inspection vision judges OK/NG; robot vision converts the recognized position into robot coordinates and carries it through to the actual pick.
AT A GLANCE
Why robot vision is different — 3 reasons
Inspection vision looks at a single aligned product up close and outputs OK/NG. Robot vision works under different conditions altogether. The numbers show the difference.
01 · Shooting distance
Clearance from the robot
To leave room for the robot arm to move, the shooting distance grows to at least 650 mm. The lighting method and setup technique differ from inspection vision.
02 · Shooting range
Many products captured across a wide area
It tracks a range of 200–400 times the part area. In a wide FOV full of irregular, overlapping and flipped parts, it has to find a position it can pick.
03 · The result is robot coordinates
Found and processed within 11,250× the part area
Pick by precise coordinates
It outputs X/Y/θ coordinates in mm, not OK/NG. That these numbers are actually achievable is the difference of robot vision.
Which one is the vision you use today? — 6 diagnostic checks tell you quickly.
Diagnose your vision now →WHY
Why the two kinds of vision must be told apart
Both are called "vision," but the result they produce is different. Inspection vision outputs an inspection result (1/0). Robot vision outputs a physical position (X/Y/θ). Because the kind of output data is different, one vision system cannot play both roles.
Inspection vision
Outputs an inspection result — 1 or 0
It analyzes the image and judges pass or fail. The result is 1 or 0, a physical signal. It cannot tell the robot where to move.
Robot vision
Outputs a physical position — X / Y / θ
It computes the physical position of the part from the image. The result is X / Y / θ coordinates — used directly for the robot to move to that position. It is position data, not an inspection result.
Item-by-item comparison and the technical differences — core technology / robot vision →
HOW IT WORKS
How robot coordinates are made from vision
From the camera recognizing the part to the robot receiving its pick command — a 4-step processing flow.
DIAGNOSTIC
Is the vision you use today the right type for a robot?
Check with 3 questions. If even one is YES, inspection vision has most likely been slotted into your robot pick-and-place flow.
- 01
Is the only result your vision outputs a 1/0 (pass/fail signal)?
If YES — If there are no X/Y/θ coordinates, it is inspection vision. The robot cannot know where to move, so it cannot perform a pick.
- 02
Is the camera mounted on the robot arm, or is the shooting distance so short that it interferes with the robot?
If YES — Robot vision needs a distance of at least 650 mm to capture without colliding with the arm. Inspection vision designed for close-range capture does not meet this condition.
- 03
Are irregular, overlapping parts scattered across a wide area, and do you have to find which one to pick?
If YES — Inspection vision assumes a single, aligned product. Finding a pickable position among many parts in a wide FOV is the job of robot vision.
If the diagnosis is unclear — send us your current system information. We will review the fit with you.
MISMATCH
When inspection vision is used on a robot as it is
The mismatches we often see on the floor and their consequences. Designing with robot vision from the start avoids these problems.
-
Situation
Connecting inspection vision straight to a robot
Result — Only an OK/NG result comes out and there are no pick coordinates, so the robot has no idea where to go.
-
Situation
Passing image coordinates without coordinate conversion
Result — They do not line up with the robot coordinate system, so every run needs correction or risks a collision.
-
Situation
Inspection vision with no Recipe function
Result — Conditions have to be entered by hand every time the model changes, which makes operation hard.
-
Situation
Running separately with no integration
Result — Vision, robot and feeder move on their own, so trigger and synchronization problems keep coming back.
START
We will review whether the vision you use today suits your robot.
Tell us your current vision system, robot and part conditions, and we will check the fit with you.