In a food processing plant, hundreds of dates speed along a conveyor in the packaging phase. There’s no way a person could locate all the damaged fruit before it’s ready to be boxed. An installed machine vision system signals the damaged goods early in the process.
Packaging food, inspecting the quality of small items like weld nuts and bearings, and robots painting tractors moving along an assembly line are examples of jobs using machine vision. Systems are set to exact specifications so the likelihood of errors are greatly reduced or completely eliminated.
Human inspectors can’t possibly see all the flaws in products speeding on a conveyor at up to 500 parts per second. Automated vision gives real time alerts to signal problems so operators can make corrections.
An effective system requires integrating a company’s goals and objectives with technology that itself is a blend of distinct capabilities. The first step is knowing exactly what you want to achieve.
Planning is key when installing any automated system. Pre-planning maximizes machine vision effectiveness. Here are a few questions to evaluate a system before it’s installed.
For an inspection system, what task must be performed? Will it ensure that components are installed correctly or colors on a product have the required tone?
What’s the impact of the operating environment? Heat, humidity, and vibration affect the choice of lighting, lens, and camera.
What third party system does the inspection system need to interface with?
Go in-depth on the subject with this write-up Robotics and Vision at a Glance: The Dos, Don’ts, and Applications.
The Benefits of Consistent Quality
Industrial automation produces high quality goods because the equipment is programmable to pre-set specifications. Machine vision determines dimensions like length, height, and width. This eliminates operator variances. When inspecting food, an item cut to a specific size, or a label on a bottle, one person may check the product and approve it while another person may say the dimensions are slightly off.
Packaged items can be counted and if a box has five widgets instead of six then an error message is sent.
A vision guidance system used with a robot in spot welding applications ensures the components are properly affixed. In aerospace, for example, airplane welds must be perfect.
Vision systems zoom in on small parts and show precise locations.
Lighting, Lenses, and Cameras
The basic principle of machine vision is easy to understand. Light shines on a target and the light reflects back to a camera. The image is collected and data is processed. Measurements on the image are sent to a robot or other system and an item that doesn’t measure up is flagged. The process happens in a blink of the eye.
The components used in a system are like those found in a photography studio. It begins with lighting, lenses, and cameras. Proper lighting with the right lenses and camera can work with or without robots on conveyor systems, assembly lines, and CNC machines.
Why Robots Need to See
Robots need visual perception to be flexible and handle a range of tasks like picking and sorting items. Vision Guided Robots (VGR) are being used to work safely in assisting people. The challenge comes from integrating the robot, the vision system, and real world coordinates. The lighting and field of vision challenges are described in the article Machine Vision Moves Industrial, Collaborative Robots Forward.
Multiple cameras are used with the robot for accurate movements. This includes stereo vision, overhead, and end-of-arm mounted cameras that can move with what’s termed six degrees of freedom in a 3D space. Robotic vision guidance systems scan objects in 2D and 3D environments.
Vision-guided robots are being developed to move freely around people and physical barriers in factories and in places like hospitals. In manufacturing, the robots can assemble parts to one-thousandth of an inch. Once a system is operating well, expect years of productivity without the fatigue or error factors found with people.
Tracking and Tracing
Vision systems can serialize specific parts. In the automotive industry, door panels in cars and trucks will soon have specific codes. If something goes wrong for the consumer who says the vehicle is a lemon, the tracking and tracing will aid in failure analysis.
Companies that investment in vision systems discover operational cost savings by having fewer errors and resulting in greater profitability. Educate yourself on industrial applications for machine vision and the latetst news by visiting visiononline.org and clicking on the Education page.