Machine vision
Overview
Machine vision combines the picture-taking ability of a camera with the processing power of a computer to make decisions about the position, quality, and completeness of a manufactured product and its parts. Most machine vision systems include a library of software tools that perform different types of inspections and let you incorporate multiple inspections from captured images.
Machine vision excels at quantitative measurement of a structured scene because of its speed, accuracy, and repeatability. For example, on a production line, a machine vision system can provide 100% coverage, inspecting hundreds or even thousands of parts per minute. When it comes to assuring product quality, machine vision is available full-time and never takes a coffee break. Machine vision can make decisions in milliseconds, enabling you to run production lines faster. And it can accommodate a wide range of products and applications.
Today’s consumers expect the highest quality and unconditional safety in all aspects of their lives. At the same time, the regulatory climate in most industries makes robust, reliable product inspection a must-have. For example, regulations have spurred demand for machine vision systems in the food and beverage industries, and companies in the pharmaceutical and medical device industries must address consumer safety and government regulations while guarding against counterfeiting.
The alternatives to machine vision for inspection (low-cost photo sensors and human inspectors) have their drawbacks. Low-cost photo sensors can perform only a limited number of simple tasks, such as position verification, counting, and color checks. In addition, because of their rigid mounting setup, they can only provide go/no-go decisions. Human inspectors can get bored or distracted and may not be able to keep up with production line speeds. Moreover, some manufacturing environments may not be safe for human inspectors.
What can machine vision inspection do?
The versatility and industry range of machine vision inspection can be seen in the following list of typical machine vision inspection tasks:
Determine the position of an object, for example, to verify proper label placement
Ensure the package integrity of medical products, such as checking that medicine vials are fully closed and secured with tamper-proof seals
Verify that an object’s properties meet quality standards, such as the position and volume of glue beads used in automobile transmissions
Inspect manufactured goods and identify flaws, such as surface scratches, bent needle tips, and incomplete solder traces
Count items, such as pills in a blister pack, bottles in a case, and components of a kit
Check on characteristics of completed assemblies, such as the inclusion of consumer safety inserts and the completion of an assembly operation
Detect tool wear in machining operations before parts are out of spec
Measure dimensions on a microscopic level
By looking at a few specific examples of applications in the food, beverage, pharmaceutical, medical device, and automotive industries, you can understand how machine vision can support the manufacturing processes.
What's the difference between Machine vision cameras and regular cameras?
Most consumer cameras have rolling shutters, meaning each row of the image is captured at a slightly different time. This distorts the images of moving objects in a particular way but gives sensors with low noise and high sensitivity to light. Many machine vision cameras have “global shutter” sensors that can, or always do, expose all pixels simultaneously. This gives an image undistorted by motion, which is better for precise inspection.
Consumer video cameras have compressed output formats designed to store many images on portable memory cards while the camera is hand-held. Machine vision cameras are usually connected to a computer via a fast interface (1 Gbps Ethernet, 10 Gbps Ethernet, 5/10 Gbps USB 3, or even faster camera-specific interfaces). The camera sends uncompressed images (sometimes raw Bayer, sometimes RGB) to the computer, which is responsible for handling storage and display.
Consumer video cameras shoot at pre-defined frame rates designed to synchronize with a TV set at one of a few standard frequencies (e.g., 24, 25, 29.97, 30, 48, 50, 59.94, 60 FPS). Machine vision cameras often include an external trigger input so that they can be triggered every time a relevant event happens (e.g., a package goes by the camera on a conveyor), even when those events occur at irregular times.
Machine vision cameras often come with an interface library designed to allow the user to write code that fully controls the camera and reads its image data.