Robotic vision is the practical application of computer vision providing semi-autonomous machines and devices with a limited visual capacity as needed in industrial and manufacturing settings. Emulating, but not duplicating human vision, machine vision allows for the recognition, identification and assembly of parts with minimal human interference.
Related Categories
Industries such as pharmaceutical, food processing, electronics, textile, medical, automotive, recycling, construction, aeronautics and automation industries commonly utilize robotic vision to aid in repetitive tasks such inspection, code scanning, defect recognition, sorting, guiding. While robotic vision does not provide the inference and comprehension of a human workforce, it does provide savings, increased production speed, reliability and quality in many applications. Automation equipment such as this is also employed in hazardous environments, though it is important to consider the capabilities and limitations of a specific device with regards to environmental exposure. In these and other applications robotic vision is used for both object recognition and navigation. While the former entails the interpretation of the data present in visual imagery, the latter uses positioning markers in conjunction with this extrapolated data to offer guidance and assist in the placement of crucial components such as electronic circuitry. These robotic vision systems are highly integrated and complex, but the simple addition of software and a camera will often suffice to provide at least low-grade vision capabilities.
Robotic vision is just one component within a larger work cell, another name for the equipment in its entirety. Just as with a human workforce, vision is a key aspect as it informs the actions of a device in response to surroundings. Machine vision systems can be located either at the feeder point or at the end effector or arm of the robot depending on its role and the order of processing required for a specific application. Systems may vary significantly, but most begin with a vision sensor. This device triggers a camera or other image making device when needed. A frame grabber converts the data from this image and stores it in a computer memory system. Specialized software then uses advanced software to interpret this data and trigger the necessary robotic action. Robotics with high processing speeds can provide real-time data that offers high accuracy and significant gains in production time. Though robotic vision is best equipped to perform repetitive work that is narrowly defined, some vision systems are outfitted with an accessible interface allowing users to switch from program to program. This alters the recognition and evaluative cognition of the robotic device as needed for the completion of different tasks.
