Machine Vision: Why Your Industrial Robot Needs Eyes
Industrial robots safely and efficiently perform tasks that are either too repetitive, difficult or hazardous for human workers and do so in a way that enhances production by eliminating bottlenecks and increasing product quality, but without “eyes” to see what they’re doing, robotic applications can be limited. Since a robot can only perform a task within the range of motion for which it has been programmed, a human worker is often required to precisely set up the robotic workspace to ensure that objects are positioned in a way that the robot can “blindly” locate them and perform the associated tasks that require intelligence. However, giving your industrial robot the gift of sight via machine vision can greatly expand their capabilities. When guided by a machine vision system, industrial robots can independently use their “eyes” to detect, locate, select and inspect objects within the defined workspace.
What is Machine Vision and What Can It Do?
Machine vision incorporates several components, including sensors and a camera to collect image data, analytics software and a robot controller, which are discussed in more detail below. The combined technologies provide visual feedback that allows a robot to complete a variety of tasks using object, pattern, symbol and signature recognition. These machine-vision capabilities expand the ability of robots to complete more complicated tasks in industrial manufacturing applications.
Vision-enabled robots may be found performing inspection, quality control, assembly, pick and place and other jobs in a variety of industries. For instance, they may guide assembly or welding robots or verify the orientation of parts in the automotive industry, analyze a particular characteristic of a part being manufactured on an assembly line and then inspect it to determine if the part meets quality standards, examine pharmaceuticals for contaminants or defects acquired during the manufacturing process, monitor a beverage bottling line to ensure that caps are correctly applied, read bar codes for stock management and pick objects based on the position of targets. Machine vision-enabled robots can do all of these activities faster and with fewer errors than human workers, increasing the efficiency and profitability of the process, while boosting quality and yield and reducing defect rates.
The ability of a robot to “see” provides more value in these types of applications than you might imagine. Think of a pharmaceutical labeling task in which a machine-vision-guided robot applies labels without any errors, ensuring regulatory compliance and preventing product recalls and the associated costs. Another value add is the ability to catch bad parts before they are built into larger assemblies, which not only increases quality and yield, but also eliminates the cost associated with scrapping larger assemblies later in the process. Another example of additional value is the ability to detect and remove misaligned parts from an assembly line before they cause a machine to jam, preventing not only maintenance costs, but also the high price of the associated downtime and lost productivity.
How Do I Give My Robot “Eyes?”
Machine vision systems typically require a sensor and camera that are used to capture the necessary images, including parts, bar codes, symbols, objects or obstacles; a processor of some sort such as a PC loaded with software programs that contain machine vision algorithms and artificial intelligence for analyzing and processing the images; and an output that uses the collected and analyzed data to communicate with and direct the robot’s next steps by sending information to a robot controller.
To illustrate how it works in an industrial application, let’s look at the process of inspecting a part on an assembly line. A sensor detects the presence of a part, which then triggers the camera to capture the image. The analog image is then digitized, stored in the computer’s memory and sent to software for processing. Using specialized software that is equipped with intelligence and analytics, the image is searched for defects as determined by pre-programmed characteristics. Based on the analysis, if the part passes inspection, it continues down the line and if it fails inspection, the robot is directed to remove it from the line.
While many of today’s advanced industrial robots come equipped with machine vision systems, it is also possible to add machine vision to an existing robot with the addition of sensors, cameras and machine vision software components. Please note that when taking this route, it’s essential to ensure that the machine vision system is compatible with the existing robot and control programs.
When adding machine vision to an existing robot, there are two ways to do so: the addition of robotic vision sensors or the addition of a complete robotic vision system. A robotic vision sensor provides an integrated camera and controller and is designed for easy installation and operation. Machine vision sensors are best suited for simple tasks, such as part inspection, that require only a pass or fail response.
Robotic vision systems are more complex and include a camera that is mounted. Cameras used in robotic applications are not required to be mounted on the robot but can be stationary above the robot’s area of interest. These cameras are connected to a PC equipped with a software package that processes and analyzes the images captured by the robot’s camera. Robotic vision systems are recommended for more critical and/or complex tasks such as assembling products, welding, critical inspections for regulatory compliance and unloading pallets.
Banner, Cognex, Datalogic, Teledyne Dalsa and more.
Whichever route is chosen, giving a robot “eyes” will increase its flexibility, efficiency and ability to complete tasks with less human intervention, resulting in greater productivity, yield, quality and profits for your facility. To learn more about machine vision and the necessary components, please contact an expert at John Henry Foster.
Want to see how much you can save by purchasing a robot?
Our Robotic Savings Calculator will help you estimate the three-year cost savings of implementing a robotic solution from John Henry Foster. Savings can be seen in cost of employee vs robot, savings in cost per production unit, and savings in operating cost per hour.