MACHINE VISION GLOSSARY

What is Machine Vision?

Machine vision is the use of cameras, sensors, and image processing algorithms to enable machines to automatically see, inspect, measure, and make decisions in industrial environments.

DEFINITION

A complete definition of machine vision

Machine vision is the engineering discipline and set of technologies that enable machines to interpret visual information from the physical world and act on it. A machine vision system typically consists of one or more industrial cameras, controlled lighting, optical lenses, image capture hardware, and processing software that analyzes the acquired images to perform inspection, measurement, identification, or guidance tasks. In manufacturing, machine vision replaces or augments human visual inspection with automated systems that are faster, more consistent, and capable of operating continuously without fatigue.

The fundamental operating sequence of a machine vision system follows four stages: image acquisition, image processing, feature extraction, and decision output. First, a camera captures an image of the target object under controlled lighting conditions. The image processing stage applies algorithms to enhance, filter, and segment the image. Feature extraction identifies relevant characteristics such as edges, shapes, textures, colors, or patterns. Finally, the decision stage compares extracted features against predefined criteria or learned models to produce an actionable result, such as pass/fail, dimensional measurement, classification, or position coordinates for robotic guidance.

Machine vision is deployed across a broad range of industries and applications, including quality inspection in automotive and electronics manufacturing, food safety and contamination detection, pharmaceutical compliance, logistics and warehouse automation, agricultural produce grading, infrastructure monitoring, and transportation safety. Modern machine vision systems increasingly incorporate artificial intelligence and deep learning to handle complex, variable inspection tasks that traditional rule-based algorithms cannot reliably address.

SYSTEM ANATOMY

Key components of a machine vision system

01

Cameras & Sensors

Industrial cameras capture images of parts, products, or scenes at high speed and resolution. Common sensor types include area-scan, line-scan, and 3D sensors, each suited to different inspection geometries and throughput requirements.

02

Lighting

Controlled illumination is critical to machine vision accuracy. Techniques such as backlighting, ring lighting, structured light, and diffuse dome lighting ensure consistent contrast and minimize shadows, reflections, and ambient interference.

03

Optics & Lenses

Lenses determine the field of view, working distance, and magnification of a machine vision system. Telecentric lenses eliminate perspective distortion for precision measurement, while standard C-mount lenses serve general-purpose inspection.

04

Image Processing Software

Software algorithms analyze captured images to extract features, detect defects, measure dimensions, and classify objects. Processing pipelines typically include pre-processing (filtering, thresholding), feature extraction, and decision logic.

05

AI & Deep Learning

Deep learning models, particularly convolutional neural networks (CNNs), enable machine vision systems to detect complex, variable defects that rule-based algorithms cannot handle. Transfer learning allows rapid model training with limited sample data.

06

Communication Interfaces

Machine vision systems communicate with PLCs, MES platforms, and robotic controllers via industrial protocols such as GigE Vision, OPC-UA, Ethernet/IP, and Profinet. Low-latency communication is essential for real-time reject and sorting decisions.

KEY DISTINCTION

Machine Vision vs. Computer Vision

Computer vision is the broad scientific field concerned with enabling computers to derive meaningful information from digital images, video, and other visual inputs. It encompasses academic research and applications across domains including autonomous driving, medical imaging, augmented reality, facial recognition, and satellite image analysis.

Machine vision, by contrast, is the industrial application of computer vision. It specifically refers to the use of imaging-based inspection and analysis in manufacturing and production environments. Machine vision encompasses not only the software algorithms but also the complete engineered system: cameras, lighting, optics, enclosures, mounting hardware, and integration with factory automation infrastructure.

Summary

Computer Vision

The science of teaching computers to interpret visual data. Broad academic and cross-industry discipline.

Machine Vision

The industrial application of computer vision on the factory floor. Includes the full engineered system, not just algorithms.

Relationship

Machine vision is a subset of computer vision, focused specifically on manufacturing, quality control, and industrial automation.

TERMINOLOGY

Key machine vision terms

AOI (Automated Optical Inspection)
A machine vision technique used primarily in electronics manufacturing to inspect printed circuit boards (PCBs) for defects such as missing components, solder bridging, and misalignment. AOI systems use high-resolution cameras and pattern-matching algorithms to compare captured images against a known-good reference.
ROI (Region of Interest)
A defined sub-area within a captured image where the vision system concentrates its analysis. Limiting processing to an ROI reduces computation time and improves accuracy by excluding irrelevant background regions.
Edge AI
The deployment of artificial intelligence inference directly on local hardware (such as an industrial PC or embedded GPU) at the point of data collection, rather than in the cloud. Edge AI eliminates network latency and ensures data sovereignty in production environments.
Inference
The process of running a trained machine learning model on new input data to produce predictions or classifications. In machine vision, inference occurs each time a captured image is analyzed by a deep learning model to detect defects or classify objects.
GigE Vision
An industry-standard interface protocol for industrial cameras that uses Gigabit Ethernet for high-speed image transfer. GigE Vision supports long cable runs (up to 100 meters), making it suitable for large-scale factory installations.
Deep Learning
A subset of machine learning that uses multi-layered neural networks to learn hierarchical feature representations directly from data. In machine vision, deep learning excels at detecting complex, variable defects that are difficult to define with hand-crafted rules.
Transfer Learning
A machine learning technique in which a model pre-trained on a large general dataset is fine-tuned on a smaller, domain-specific dataset. Transfer learning significantly reduces the amount of labeled training data and time required to deploy accurate machine vision models.
Point Cloud
A three-dimensional dataset composed of individual points in space, each defined by X, Y, and Z coordinates. Point clouds are generated by 3D sensors (structured light, laser triangulation, or time-of-flight) and are used for volumetric measurement, surface profiling, and robotic guidance.
Structured Light
A 3D imaging technique that projects a known pattern of light (stripes or dots) onto an object and measures the deformation of the pattern to calculate surface geometry. Structured light systems are widely used for high-resolution 3D inspection in manufacturing.
OPC-UA (Open Platform Communications Unified Architecture)
A platform-independent, service-oriented communication protocol for industrial automation. OPC-UA enables machine vision systems to exchange data securely with PLCs, SCADA systems, MES platforms, and cloud services without vendor lock-in.
PLC Integration
The connection of a machine vision system to a Programmable Logic Controller (PLC) via digital I/O signals or industrial fieldbus protocols. PLC integration allows vision inspection results to trigger real-time actions such as rejecting defective parts, stopping a conveyor, or signaling a robotic arm.
MES (Manufacturing Execution System)
Software that manages and monitors work-in-progress on the factory floor. Machine vision systems feed inspection results, pass/fail data, and quality metrics into the MES to provide real-time production visibility, traceability, and statistical process control.

Frequently Asked Questions

01
How much does a machine vision system cost?
02
What industries use machine vision?
03
What is the difference between 2D and 3D machine vision?
04
How accurate is machine vision inspection?
05
Do I need AI or deep learning for machine vision?
06
How long does it take to deploy a machine vision system?
OspreyX Logo
From Understanding to Implementation

Ready to explore machine vision for your facility?

Talk to our engineers about your inspection challenge. We will assess feasibility, recommend the right approach, and show you what production-grade machine vision can achieve.

High Accuracy from Day One

Solution Design

Transparent Pricing & Specs

Live Demo

Get Your Machine Vision Assessment