Our Sites

Vision-guided robotics cement their place in the press shop

How vision systems enhance robotic automation performance

 a vision system transmits data to a robot

With a camera positioned above the centering station, a vision system can locate the blank and transmit the X and Y coordinates and rotation data to a robot, which then picks up the blank and transfers it to the press accurately.

During the last 40 years, industrial robots have been integrated within almost all manufacturing processes as their flexibility, accuracy, and reliability have become evident. Advancements in robotic technology are extending their functionality and improving their performance. Often a robot is expected to do more than move from programmed position A to position B; it must operate “smarter” and act upon other inputs, such as vision systems and force feedback.

The applications for integration of vision systems and robots are almost endless, but the most common is vision guidance for the robot.

The key elements are:

  • Camera. A variety of devices are available, including single-camera systems, multiple-camera systems, line scan cameras, contact image sensors, and 2-D and 3-D vision.
  • Processing unit and associated software. This could be a PC or a smart camera with the power to process image data and send the required positional offsets to the robot.
  • Lighting. This is a key element for the installation of a vision-guided system. Consistent illumination is probably the most important factor.
  • Communication. This comprises all the associated elements to provide communication with the robot, as well as support structure for the cameras and lights.

The main reason the use of vision systems for part location is surpassing traditional, mechanical methods such as fixtures, gravity tables, and programmable mechanical centering tables is the flexibility they provide. For example, there are ways to ensure repeatable part position using mechanical devices, but if the shape of the parts or the number of parts varies significantly compared to the original specifications, chances are that the system will need to be changed. A vision system can accommodate large variations in shape and size so that this limitation disappears.

Process and Applications for Vision-Guided Robotics

Vision-guidance systems generally are positioned at two main areas in an automated stamping process: front of line in destacking and centering/loading in the first press and end-of-line racking and loading of stamped parts. During between-press transfer, the part typically is in a positive location in the die or transfer nest, so vision systems are not required.

Front of Line. There are two phases within front-of-line vision guiding. Typically they are considered separately for cycle-time purposes, but they also can be combined into a single operation.

  • Destacking, or picking parts from one or multiple stacks to feed the line
  • Centering, or loading the first press accurately to prevent die damage

Compared to hard automation, robotic destackers for unloading blanks have significantly improved process flexibility. Robots can pick blanks from multiple stacks and accommodate variation in position on load tables (see Figure 1).

With minimal programming changes or recipe-based external feedback, they also can drop the blanks at different locations. With the additional capabilities a vision system brings, the robots can be guided to the correct location of the blank stacks. Typically, a vision-guided robot system uses a camera located above each of the destack tables.

The vision software is programmed to identify and locate the blank stacks using geometric vision tools. The camera takes a picture of the destacking table area and locates the position of the stacks (see Figure 2). This information is then transmitted to the robot. The robot uses the X and Y coordinates and rotation data from the vision system to automatically adjust its path and pick the blanks from their respective stacks.

The operator can now teach the robot path (operational trajectory of the tool) with respect to the tooling and initial stack location without having to worry if the upcoming stacks are not located in the identical position. The robot compensates using offsets from the vision system and can pick from any new stack location on the destacking table.

Robotic destackers pick blanks from multiple stacks

Figure 1
Robotic destackers can pick blanks from multiple stacks and accommodate variation in position on load tables.

Once robots have de-stacked the single blanks, they typically are dropped directly onto a centering station or moved there by conveyor. The purpose of a centering station is to accurately adjust blank position so that it can be loaded into a press (see lead image)

With a camera positioned above the centering station, a vision system can be programmed to locate the blank and then transmit the X and Y coordinates and rotation data to a robot. The robot now can use this information to pick up the blank, implement offsets, and place it into the press accurately. All operators are required to do is teach the robot its pickup point related to the tooling and blank once. Then the vision system communicates positional variation of all upcoming blanks to the robot, which will compensate for these offsets.

Before vision-guided robots, mechanical devices with numerous adjustable guides were used. In today’s stamping operations, blanks are not just rectangular; they come in a variety of shapes and sizes. Most press lines process various part types. Without vision guiding, it was costly to build complex centering stations that could accommodate adjustability. Maintenance costs were cost-prohibitive too. With a vision-guided robot, adjustability can be achieved reliably and accurately.

End of Line. Many end-of-line systems are referred to as pack-out systems. These usually have long conveyors so that the formed parts move along the conveyor at a speed slow enough for operators to be able to retrieve the parts and load them into racks. Technology has increased the rate at which press lines can run; however, the higher speeds create problems for the personnel who are racking these parts.

A solution to this problem is to use robots with vision guidance. A typical robotic end-of-line system has a conveyor with a camera mounted above it at the load end. As parts move along the conveyor, a picture is taken and the part location is sent to the robots. Alongside the conveyor, robots pick up parts and load them into racks placed within their reach. A secondary vision system above the racks identifies and communicates the location of key repeatable geometric features of the rack that can be reliably identified by the vision camera and provide consistent positional information to the robot. This is required usually because of the construction and extensive wear of part racks. The robots can rapidly pick up the part on-the-fly as the conveyor is moving, and then insert it into the rack.

In a simplification of this concept, an unloader robot retrieves a part from a press and hands it over directly or through a nest to a racking robot. The vision system is located above each rack—typically two—and identifies and communicates the location of the key loading features of the rack.

Cost justification for the vision system is the same—rack construction and wear. This approach is efficient and completely eliminates operator intervention.

Lighting Essential to Reliability

As was mentioned at the beginning of this article, one key factor to consider is lighting. Good lighting is essential to ensure a reliable and repeatable vision system.

Typically, lights are mounted at a height on overhead structures that allows ambient light into the work area. The light level variations can produce inconsistencies in the image acquisition and processing. It is quite hard for an artificial light source to supersede potential ambient light, so one option is to use a shroud around the work area. This is not always possible, however, because of the location of existing equipment or the robot’s work envelope requirements. Another option is to use near-infrared light, the downside of which is its cost and possible reduced efficiency.

The newest generation of robotic centering stations use line scan cameras or contact image sensors (see Figure 3). In line scan camera systems, powerful lights are installed in the proximity of the part, rather than the camera, to diminish external lighting.

vision-guided robot system

Figure 2
Typically, a vision-guided robot system uses a camera located above each of the destack tables.

Lights and lenses are incorporated in the body of contact image sensors. In addition, these systems maintain a very low profile so an overhead structure is not required. The low profiles also are minimally intrusive into the robot’s work envelope.

In Vision-Guided Robotics, Flexibility Is Key

A vision-guided robotic system’s advantages can be summarized in one word: flexibility. Besides the requirement to build robot tooling for each part or part family, the only other consideration for part introduction to the press is programming. The new part is introduced to the vision system’s field of view; the robot learns geometric features so that the camera can identify the part; and the robot is programmed to pick and drop the part.

From that moment the part can be run on the line. This applies to destacking, centering, or racking as a general idea. There is no need to rebuild or replace hardware extensively, which also would have to be reprogrammed. The material and engineering time savings equates to significant cost savings.

Figure 3
In line scan camera systems, powerful lights are installed in the proximity of the part to eliminate external lighting influence.

About the Authors

Abel Elias

Applications Specialist

1250 Brown Road

Auburn Hills, Mich. 48326

Enrique Pano

Press Automation Manager

201 Westcreek Blvd.

Brampton, ON L6T 5S6 Canada

905-460-3347