What is Machine Vision?

Machine vision is a technology where computer systems process and analyse images to make decisions or control processes. It is widely used in industrial automation to perform tasks faster, more accurately and more consistently than possible with the naked eye. Learn more about the technical aspects on our technology and integration page.

General benefits of Machine Vision

Machine vision systems offer benefits such as:

Speed: images are analysed in real time for immediate feedback.
Accuracy: small deviations and defects are detected reliably.
Consistency: the same criteria are applied consistently without human variation.

For advanced applications such as AI and deep learning, which further improve accuracy and efficiency, we refer to our AI & Deep Learning page.

Standard components of a Machine Vision system

A machine vision system consists of different parts that work together to create and analyse images:

Camera: captures images of the object or process.
Optics (lenses): determines the sharpness and field of view of the camera.
Lighting: provides the right illumination to make details visible.
Software: processes and analyses the images to make decisions or trigger actions.

For more information about the brands and components used, you can visit our brands & components page.

How does a machine vision system work?

Every machine vision system goes through the same four steps for each product, often within a few milliseconds:

Capture: a sensor or the machine sends a trigger, the lighting fires and the camera takes an image of the product.
Process: the software locates the product in the image and measures, reads codes or detects defects, using fixed rules or a trained AI model.
Decide: the system compares the result with the requirements you set and passes or rejects the product.
Communicate: the result goes to the PLC, robot or SCADA system, for example over Profinet, Modbus or EtherCAT, so a rejected product is removed straight away.

Types of machine vision systems

The right type of vision system depends on the task, the line speed and the number of cameras:

Smart camera: camera, processing and software in one compact housing. Ideal for a single inspection point, such as code reading or a presence check.
PC-based system: several cameras on an industrial PC, with the computing power needed for complex inspections and AI models.
2D and 3D vision: 2D vision looks at shape, colour and text; 3D vision also measures height and volume, for example to detect deformation or check a fill level.
Line scan cameras: build the image line by line and suit continuous webs such as paper, film, textiles or metal.

The software is the heart of every system. Read more about off-the-shelf and custom options on our page about machine vision software.

Examples of applications

Machine vision is used in many sectors, for example for:

Quality control of products and parts
Positioning and guidance of robots or machines
Identification and verification of labels, barcodes and datamatrix codes
Inspection and defect detection in production processes

Would you like to know more about specific applications such as label control, color inspection or retrofit? Then visit our OCR & Label control, Color inspection and retrofit pages.

Would you like to know more about specific applications or technical details? Then visit our technology and integration page or our services.

Machine vision vs computer vision

The terms are often used interchangeably. Computer vision is the broad field in which computers learn to understand images, from face recognition to self-driving cars. Machine vision is its industrial application: cameras, lighting and software that inspect, measure and guide products in a production environment.

You will also hear industrial vision, vision technology, image processing or simply vision systems. They all share the same goal: using cameras to decide reliably, and faster than the human eye, whether a product is right.