Machine Vision Systems for Robotics
Vision technology has become a critical component for many robot applications, enabling robots to be deployed into new areas. Over the years the technology has matured becoming very reliable, with higher performance and pricing has dropped dramatically. In the process vision has become accepted and embraced by many companies who see the value it brings to their operations; value in the form of improved product quality, more manufacturing flexibility and higher throughput.
Yaskawa Motoman has partnered with the industry leaders in vision technology. This allows us to deliver the very best integrated solutions to our customers. Learn how our robotic vision systems can be applied in your operations today.
While giving robots eyes doesn’t make them self-aware, it does enable them to perform simple operations in ways that dramatically improve their performance. For example, robots guided by vision can locate parts to be picked up, determine where to apply a weld, inspect parts that have been assembled, determine where to place a part. The possibilities are endless.
Giving robots eyes enables them to perform increasing complex operations in ways that dramatically improve their performance. For example, robots guided by vision can locate parts to be picked up, determine where to apply a weld, inspect parts that have been assembled, determine where to place a part. The possibilities are endless.
- Part Location
Presenting parts to robots varies widely from application to application. In the past parts needed to be precisely fixtured so that the part is always in the same position and orientation. This is expensive, time consuming and does not allow for flexibility when the part type or size changes. Our MotoSight products can help locate the parts so that the robot can then pick them up precisely and consistently. This includes parts that are loosely fixtured, parts that are presented on a conveyor and in some cases parts in a box or bin.
- Part Identification
Traceability and genealogy of parts is crucial to many manufacturing or production processes. Other processes simply require parts to be identified for a specific operation such as assembly or kitting. Our MotoSight products can help identify and validate parts based on part geometry (shape and size), color, barcode, markings and other features.
- Part Sortation
Many manufacturing and production processes require part to be sorted. Our MotoSight products can help select and sort parts based on part geometry (shape and size), color, barcode, markings and other features.
- Flexible Feeding
Flexible feeding is an innovative way to present parts to a robot. Parts are fed out of a bulk hopper and onto a feed surface. A feed surface may be a solid indexing or vibratory surface of a conveyor belt. A camera mounted above the feed surface located the part and determines its orientation. The robot then picks the parts and performs the necessary operation. If the part size or type changes, you need only make minor changes to the vision locator tool and perhaps the gripper fingers. Flexible feeding is ideal for production processes requiring multiple part types or a family of parts. Our MotoSight products are ideal for flexible parts feeding.
Operations often require validating if a part or a feature on a part is present or absent. If the part/feature is present the operation continues. If the part/feature is absent, the exception can be caught and managed. Our MotoSight products can help determine presence of a part or feature based on geometry (shape and size), color, markings and other features.
- Error Proofing
Error proofing is similar to Presence/Absence. Preventing a manufacturing error before it occurs or detecting an error early in the process can save companies from shipping faulty or defective product. Detecting holes present, threads on a bolt, caps on a bottle, part in a fixture, if a part is oriented correctly are all examples of error proofing. Our MotoSight products can be applied to error proofing in virtually any application.
Shipping faulty or defective products can be devastating to a business. Machine vision has a long history in part inspection, ensuring consistent and predictable quality. Whether it's gauging a part, detecting presence of a feature or error proofing, our MotoSight products have all the necessary inspection and data analysis tools for part inspection.
Typical vision systems consist of structured lighting, a camera with lens, a processor (sometimes embedded in the camera) and vision software. Structured lighting ensures the image the camera sees is consistent and doesn’t vary with the environment. The lens provides focus and how large an area the camera sees. The camera captures the image and sends it to the processor. The processor converts the image to digital signals. The vision software interprets the digital signals and provides the robot with useful information.
- Augmenting Vision Systems
Vision systems can be used in conjunction with a variety of other perception tools. This includes line-of-sight laser scanners, time-of-flight sensors, flash imaging LIDAR and, even Microsoft Kinect sensors are being used to augment vision systems. Additional sensing can provide a depth and feel dimension to traditional vision capabilities.
- Environmental Perception
All of the above are examples of robot systems using a vision system to perceive and respond to the environment around the robot. As vision technology advances and is augmented with other sensing technology the ability to deploy robots into a wider range of applications increasing. Now we can detect randomly positioned boxes in 3D space and depalletize these items and process them correctly. We continue to work with more complex and difficult applications, like bin picking, that have a higher degree of part variability.
- Camera Mounting
Cameras can be mounted to the robot near the end-of-arm tooling. In this example the camera moves with the robot. This can provide much better resolution (a closer look) of the part, but it also means the robot can't do other things while a picture is being taken.
Cameras can also be fixed mounted above the workspace. This might be over a particular work area, over a moving conveyor belt, or over a flexible feed surface.
Cameras can also be fixed mounted (up, down, horizontal or at a unique angle) within the workspace of the robot. The robot can them place parts in front of the camera, or hold parts in front of the camera.
- 3D CAD Matching for Bin Picking
- Motoman MH5L robot with Flexible Parts Feeder
- Motoman HP20 robot applying sealer to automotive underbody
- Motoman UP20 loads and unloads a drill machine
- Motoman HP165 picks castings from a bin
- Motoman HP20 Bin Packing Wheel Covers
- Motoman SIA20 3-D Vision Demo at IMTS
- Box Depalletizing demo featuring MotoSight 3D Spatial Vision
Most vision systems return information in 2 dimensions (X and Y) with rotation. The majority of vision applications can be solved with our MotoSight 2D product.
When an application requires information in 3 dimensions (X, Y and Z) also with rotation, 3D vision should be used. These applications might require our MotoSight 3D CortexVision or our new Motosight 3D BinPick
- MotoSight 2D Global Edition
MotoSight™ 2D Global Edition is a powerful vision system featuring the Cognex In-Sight® Micro camera, robot calibration routines and a proven tool set that enables users to solve complex vision applications. The system is programmed using Cognex EasyBuilder programming software. Updates to the vision program can be made directly from the DX100 controller programming pendant.
MotoSight 2D is ideally suited for general robot guidance, error proofing, part identification, part Inspection and high-speed picking. Target applications include machine loading, pick & place transfers, assembly, packaging, palletizing, de-palletizing and welding.
- MotoSight 2D with MotoPick
MotoSight™ 2D with MotoPick™ software locates randomly oriented products on a moving conveyor belt. The MotoPick software coordinates the picking responsibilities or the number of robots in the line. Using conveyor tracking, each robot picks the product and performs the necessary operation.
MotoSight 2D with MotoPick is designed for robot guidance in high-speed packaging, cartoning and case packing applications.
- MotoSight 3D BinPick
MotoSight 3D BinPick is an easy-to-use hardware/software solution that utilizes a Canon 3D machine vision system to quickly and easily recognize and pick parts that are randomly placed in bins. 3D CAD matching provides simplified, accurate part registration, allowing even complicated parts to be identified.
Single-step recognition eliminates the need for complicated programming of parts and can greatly reduce cycle times. Learn More
- MotoSight 3D CortexVision
MotoSight 3D CortexVision simplifies the use of 3D vision in robotic guidance applications. Using a single camera mounted to the robot, this powerful vision system is ideal for precise part location required for assembly, racking, dispensing, machine loading, product sorting and welding applications.
No calibration required! The point-and-click software environment and self-learning technology dramatically reduce setup time. The system takes multiple pictures and automatically refines the position of the camera to optimally locate the part.
- Technology Partner Solutions
For e-fulfillment and bin picking of packaged goods, Univeral Logic's Goods-to-Robot Cell is a flexible solution. The hardware/software package utilized artifical intelligence, 3D vision guidance, motion control and WMS/ASR interface provides a fully integrated order fulfillment solution.
- Vision Options & Peripherals
A variety of options are available for our vision systems, including lenses, lighting and camera enclosures. We also offer additional cameras for more demanding applications. Please contact us for details.
- Automation Shines at Vehicle Safety Systems Company
- Robots Help Assembler Break the Mold
- System helps prepare sintered metal parts
Vision Systems Design