Similar to its counterpart, machine vision, embedded vision is the application of hardware and technology to assist in process control and automation. Common applications include industrial settings, autonomous vehicles, and drones. Until recently, machine vision was viewed as a separate but related technology, something like a big sister. However, as computer technology evolves, the lines between embedded vision and machine vision continue to blur.
Characteristics of Machine Vision
Machine vision systems are usually built for general image processing, and use cameras or image sensors connected to a PC to capture and process images and visual information. These systems are typically used in industrial and manufacturing operations where they can identify anomalies and defects, and the item can be removed from production. Traditionally, these systems require ample space for operation but are relatively inexpensive to set up. The server-class computer has a high level of processing capabilities and software that is programmed to deliver numerous images from which information can be gleaned and subsequent decisions made. With a wide range of uses, machine vision is typically used in manufacturing, agriculture, inventory control and management, safety, product tracking and traceability, and so forth.
Characteristics of Embedded Vision
Embedded vision systems are compact and typically include a camera mounted on an image processor. These all-in-one devices can be used for applications where space is at a premium. Though technology is quickly evolving, embedded vision systems typically have limits on their image processing capabilities compared to more robust yet sizeable machine vision systems. Because cameras and image processors are usually tailored to each embedded vision application, these systems require higher upfront costs. However, because of their compact size, they require less energy and result in lower operational costs.
Key Differences
The primary differentiators between current embedded and machine vision systems are the size and weight of the components required, processing capabilities, ease of integration, setup costs, operational costs, energy consumption, and applications for each system. However, with the evolution of technology, PCs used in machine vision are becoming more compact, and the image processors in embedded vision are increasing in capabilities.
Embedded Vision Applications by Industry:
- Aerospace: Enhanced vision systems (EVS) and synthetic vision systems (SVS) employ embedded vision to enhance visibility, as well as aircraft control and safety. Embedded vision systems in graphics processing units (GPUs) can also monitor aircraft engine safety.
- Automotive: This industry uses embedded vision in the development of advanced driver assistance systems (ADAS) with the purpose of designing and creating fully autonomous vehicles.
- Augmented Reality: In AR, physical and virtual images merge with the help of embedded vision systems that can provide large amounts of visual data. Augmented reality systems are commonly used in manufacturing for operational efficiencies and medicine for training and educational purposes.
- Chemical & Pharmaceutical: These highly regulated industries rely on embedded vision systems to ensure quality assurance, thereby reducing production costs.
- Consumer Electronics: Embedded vision software in smartphone cameras enables face and object recognition. From TVs, computers and video game consoles to controlling the indoor environment of our homes, embedded vision systems are more present than ever in consumer electronics.
- Defense: From surveillance systems in drones to synthetic vision systems (SVS) in aircraft, embedded vision systems can increase the capabilities of militaries and their operations.
- Drones: Embedded vision systems on unmanned aerial vehicles (UAVs) or drones make these flying devices profitable in countless applications, such as agriculture, surveying and mapping, urban planning, and more.
- Medical: Smart medical devices with embedded vision technology—such as those used in dermatology or by practitioners to monitor vital signs—can help improve the diagnostic capabilities of medical professionals.
- Robotics: Manufacturers use embedded vision in robotics to improve performance and increase functionality in industrial robots. In assembly lines, this technology can improve quality control and inspection.
- Security: Embedded vision can be used in drones and video surveillance systems to detect perimeter and other intrusions. Systems then communicate pertinent surveillance information to security team members.
- Semiconductor: Rising design and production costs can be prohibitive in the semiconductor industry. However, embedded vision systems can alleviate some of this pain by performing inspection and measurement functions.
- Supply Chain Management: This industry has the potential to benefit from embedded vision systems through automated guided vehicles (AGVs), barcode scanning, inventory tracking and traceability, and other operational efficiencies.
Sources:
AZO Materials
Embedded Vision vs. Machine Vision
AIA
Vision Online, Global Association for Vision Information
What is Embedded Vision?
https://www.visiononline.org/blog-article.cfm/What-is-Embedded-Vision/90
AIA
Vision Online, Global Association for Vision Information
Embedded Vision Systems Resources
https://www.visiononline.org/embedded-vision
Allied Vision
Understanding Embedded Vision
https://www.alliedvision.com/en/productsfolder/embedded-vision/understanding-embedded-vision.html
Forbes
What Is Machine Vision And How Is It Used In Business Today?
AIndraLabs
Computer Vision: Face & Object Recognition through Smartphone Cameras
https://www.aindralabs.com/computer-vision-based-face-object-recognition-platforms/