What Is Machine Vision?

Machine vision is a technology that allows machines to capture images with cameras, process them with image-processing algorithms, and use the results to recognize, inspect, measure, or control physical objects. It is widely used in industrial automation, quality inspection, robot guidance, medical imaging, logistics, and many other fields.

A typical machine vision system consists of an image sensor, lens, lighting, camera interface, image-processing software, and a host controller or embedded processor. The goal is not simply to take a picture, but to extract reliable information from the image and turn it into a decision or control action.

Camera Types by Function

Machine vision cameras can be broadly divided into two groups based on where the image processing is performed.

  1. Image acquisition cameras

    • Capture raw images or video from the image sensor and transfer the data to a PC or host system.
    • Image processing, detection, and analysis are mainly performed by software on the host.
    • Most industrial machine vision cameras fall into this category.

  2. Smart cameras

    • Include an internal processor, such as an SoC, FPGA, CPU, or dedicated vision processor.
    • Can perform image acquisition, character recognition, barcode reading, defect inspection, measurement, and decision logic inside the camera.
    • Often send only the result data or directly control external I/O.
    • Useful when a compact, standalone, and easy-to-maintain system is required.

GenICam

GenICam, short for General Interface for Cameras, is a standard that provides a common way to control cameras and acquire images across different camera interfaces such as GigE Vision, USB3 Vision, and Camera Link.

With GenICam, software developers can control cameras without being tightly coupled to a specific hardware vendor. Camera parameters such as exposure time, gain, trigger mode, image size, pixel format, and acquisition mode can be accessed through a standardized feature model.

Image Sensor

The image sensor converts incoming light into electrical signals and is one of the most important components in a machine vision system. CMOS and CCD sensors have traditionally been the most common types, and the right sensor is selected based on resolution, sensitivity, frame rate, dynamic range, shutter type, and application requirements.

Major Sensor Types

  • Area scan sensor: Captures a two-dimensional image at once. This is the most common sensor type used in machine vision cameras.
  • Line scan sensor: Captures a one-dimensional line image continuously and reconstructs a two-dimensional image from motion. It is commonly used in high-speed production lines, web inspection, printing inspection, and continuous material inspection.
  • ToF sensor: Measures distance using the time of flight of light. It is used for 3D sensing, object measurement, robot vision, and depth-aware inspection.
  • Other sensor types: Specialized sensors such as SWIR and hyperspectral sensors are used for material inspection, chemical analysis, agriculture, semiconductor inspection, and other domain-specific applications.

Camera Interface

The camera interface determines how image data is transferred from the camera to the host system. Interface selection affects bandwidth, cable length, latency, synchronization, cost, and system complexity.

GigE

GigE Vision is an Ethernet-based standard for industrial image transmission. It supports long cable lengths, commonly up to 100 meters, and makes it easy to connect multiple cameras using standard networking equipment. It is widely used in industrial environments.

  • Transfer speed: Up to 1 Gbps for GigE and 10 Gbps for 10GigE.

USB

USB3 Vision uses the USB 3.x interface for high-speed image transfer. It is easy to install, cost-effective, and suitable for systems where cable length is relatively short and plug-and-play convenience is important.

  • Transfer speed: Up to 5 Gbps for USB 3.0 and 10 Gbps for USB 3.1.

Camera Link is a dedicated industrial camera interface designed for high-speed and high-resolution image transmission. It is commonly used in demanding industrial inspection systems where deterministic behavior and stable performance are important.

  • Transfer speed: Up to about 6.8 Gbps in Full configuration.

CoaXPress

CoaXPress enables high-speed, long-distance image transmission over coaxial cables. It is well suited for real-time, high-resolution imaging applications that require very high bandwidth and robust cabling.

  • Transfer speed: Up to 12.5 Gbps per channel with CXP-12.

Vendors

  • Basler
  • Teledyne FLIR
  • IDS Imaging
  • Hikrobot
  • Hikvision
  • Vieworks
  • Cognex
  • Sony
  • Daheng Imaging
  • Baumer
  • JAI