Terminology · Machine vision
Machine vision
Cameras and algorithms in place of the human eye, for measurement, recognition and decision. On a production line it comes down to three jobs: appearance inspection, dimensional measurement, robot guidance.
Definition
Machine vision is an engineering discipline: an industrial camera captures an image, an algorithm extracts information from it, and that information becomes an executable decision or motion. Where computer vision studies the methods, machine vision has to turn those methods into equipment that holds up against takt time, lighting and incoming variation.
Industrial machine vision usually solves three kinds of problem: appearance inspection (is there a defect), dimensional and positional measurement (where is it, how big), and robot guidance (how to pick and place). The first two have traditionally been AOI's job; the third belongs to 3D vision.
How it differs
| Term | What it is | Where it sits |
|---|---|---|
| Machine vision | The discipline and engineering field | The outer layer; everything below sits inside it |
| Appearance inspection | A task: decide whether a surface or assembly passes | A typical machine-vision application |
| AOI | A method: decide by optical imaging and template comparison | The traditional implementation of appearance inspection |
| ACI | The next-generation term for AI AOI: the decision basis is cognition, not the imaging method | The next-generation implementation |
| Robot vision | A task: supply 6D pose for picking, assembly and dispensing | The other branch of machine vision |
Where DaoAI sits in each layer
All four product lines share one world-model foundation, so the same decision principle carries across tasks.
ACI · Auto Cognitive Inspection —— the appearance-inspection branch, with the full argument
ACI equipment and software —— 2D / 3D machines and ACI OS
Robot vision —— 3D guidance, picking and assembly
In-house 3D industrial cameras —— the imaging layer
The DaoAI World platform —— the shared foundation