Machine vision systems give vision capability to existing technologies which work on a set of rules and parameters to support manufacturing applications such as quality assurance, while computer vision refers to the capture and automation of image analysis with an emphasis on the image analysis function across a wide range of theoretical and practical applications.
What are the main differences?
Machine vision systems capture images and then process them to output a decision – it contains the camera, the lens, the processor and the software to be able to do so. Computer vision doesn’t necessarily need to capture an image as it can work off saved images, interpreting the data and producing a result/set of results.
In addition, computer vision seeks to automate tasks that the human visual system can do and also focuses on how computers can be purpose-built to gain a high-level understanding from digital images or videos. It also includes new applications which leverage advances in machine learning.
For a more detailed introduction to machine vision, you can visit our dedicated what is machine vision FAQ.
How do they work together?
Computer vision sits within a machine vision system, it is the post image capture part of the process.
What are the common use cases?
The combined technologies have applications and benefits within quality assurance, and some typical examples include:
Within the automotive industry, machine vision can be used to inspect and determine that a bead has been properly applied, ensuring the assembled parts do not leak and are fully sealed.
These ensure quality, eliminate rework, repair, and scrap, especially as more parts in manufacturing are now being bonded together with adhesives. Both 2D and 3D machine vision can be used to solve these applications. More advanced setups use special machine vision tools to solve bead applications.
An additional use case in automotive is end-of-line inspection and barcode reading. Powertrain/propulsion process inspection to ensure that the engine/transmission is correctly assembled, there are no missing parts, extra “bonus” parts, and that the clips are all in place. Engines/transmissions have many direct part marked datamatrix codes on them, these codes are read at numerous points in the process as part of the track and prace procedures. The same camera can carry out both inspections and code reading.
Food and Beverage
In food and beverage, there is a use case with the bottle cap and fill process where the system can inspect to determine if the bottle is filled and the cap is correctly applied to the bottle. This reduces scrap and waste while ensuring that the product is complete and safe.
Within the solar industry, the system can inspect the solar panel assembly process to determine that the panels are correctly built via presence/absence, location and measurement which ensures that parts being produced will work when completed and achieve maximum efficiency.
Durable Consumer Goods
In durable consumer goods (e.g. dishwashers, ovens, refrigerator, microwave) component inspection can determine that the machines and assemblies are correctly built via presence/absence, location, measurement and color identification tools.
In fastener manufacturing, the system can inspect fasteners to ensure that they are properly formed. Normally the tops and the threads of the fastener are inspected. This ensures that the parts are correctly formed which helps to determine quality and make sure that bad parts do not make it to the end customer.
Plastic Injection Molding
In plastic injection molding, machine vision can inspect and ensure that the parts being molded are fully formed. The deformed parts could be a result of short shots, meaning not enough material has been inserted into the mold and the part is deformed. This improves quality and reduces waste.