Embedded vision is the use of computer vision in embedded systems. Historically, digital image processing started and expanded in industrial and scientific applications with image processing algorithms running on powerful PCs. New powerful, low-cost and energy-efficient processors can now run image-processing applications on smaller, sometimes mobile devices without the need of a PC.
What are Embedded Vision Systems?
Typically, embedded systems are small, lightweight and low-cost computing devices that can be embedded into a larger system – for example a car, a robot, or a vending machine. When it comes to adding vision to an embedded system, you need to take into consideration the specificities of an embedded system vs. a PC-based system.
Size and weight
To build an image-processing system into a robot or a drone, size and weight are important issues. A camera for embedded vision should be very compact and lightweight. Most of the time it does not even need a housing, so a bare board design is ideal.
In an embedded vision system, you need a camera interface like MIPI CSI-2 and USB 3.0 that is compatible with embedded carrier boards, powers the camera, and takes as little space as possible. The MIPI CSI-2 interface, in particular, is gaining popularity with embedded designers.
In order to be easily and successfully integrated into an embedded system, cameras need to support embedded platforms and operating systems, primarily Linux and Linux for ARM. MIPI CSI-2 camera modules should also support the popular Video4Linux 2 standard for image transmission. OpenCV compatibility is a must since this image-processing library is the most popular in the embedded world.
Learn more about embedded vision
ALVIUM® is the combination of a unique, proprietary System on a Chip (SoC) designed for embedded and PC based computer vision and a comprehensive image-processing library (IPL).
Allied Vision implemented MIPI CSI-2 as a camera interface, creating a complete camera module with advanced image pre-processing capabilities.