The difference between vision sensors and vision systems is relatively simple:
A vision sensor performs simple inspections, such as answering yes or no to a question on a production line. A vision system performs complex tasks such as helping a robotic arm weld parts together in an automated factory.
Machine vision sensors capture the light waves emitted by a camera’s optics and work with digital signal processors (DSPs) to translate the light data into pixels that generate digital images. The software then analyzes the shape of the pixels to reveal essential information about the object being photographed.
Automated production does not necessarily refer to robots that make trucks and smartphones. Many tasks in automated factories require simple and clear vision sensor data:
Presence or absence. Is there a part in the sensor’s field of view? If the sensor answers yes, the machine vision software gives the green light for the part to be sent to its location in the production process.
Inspection. Is the part damaged or defective? If the sensor detects defects, the part is removed from production.
Optical Character Recognition (OCR). Does the part contain specific words or text? The answer to this question helps automated systems sort products based on brand name or description.
Machine vision systems use multiple sensors to perform all of these basic tasks and other more complex challenges:
Guiding/alignment: When parts require exact positioning or alignment, vision systems use sensors to identify the correct parts and place them exactly where they should be.
Code reading: Codes on packaging and individual components contain critical data that vision systems acquire in real time to sort finished products and distinguish between different parts in a production process.
Measuring: Sensors can ensure that machined parts have been cut to the correct dimensions.
3D Vision: Sensors create three-dimensional representations of parts and products. These images can be used to automate inspections and indicate to the robot arm where to pick and place parts.
Each company has to decide whether it needs simple vision sensors or more advanced vision systems. Vision sensors are designed to be easily installed and implemented so that plant personnel can install and configure them without much external assistance. If the vision task requires only a simple good/bad decision, vision sensors will meet the company’s needs perfectly.
Conversely, vision systems require greater expertise and a considerable investment of time and money for configuration, installation and training. Companies often turn to third party integrators who have solid expertise in vision system installation.
Each company in the machine vision industry has its own definition of the difference between sensors and machine vision systems. Cognex, for example, develops vision sensors that perform specific types of tasks, such as quality control for the food industry. Our vision systems combine advanced software with industrial-grade cameras to enable a wide range of industrial automation applications.
To differentiate between vision systems and sensors, you have to imagine hundreds of beer bottles on a conveyor belt in a bottling plant. A vision sensor can ensure that each bottle has a cap. If the cap is present, the bottle is approved and sent to the packaging stage, where another sensor ensures that each pack has six bottles.
But the company may want to detect caps that are deformed beyond a certain angle. It may also want to make sure that several varieties of beer have not been accidentally mixed in the six-bottle packs. A vision system will then be better suited to its needs.