Robot Vision: How to Maximize the Flexibility of Robot Technology

Robot Tech

Robot Vision maximizes flexibility in automation

Robot Vision: How to Maximize the Flexibility of Robot Technology

Catherine Bernier
Content Manager
,
Content for Cobot

Robot vision is providing a lot more automation possibilities in many applications and industries. Let’s discuss how robot vision can maximize your automation flexibility.

Robotic technology is revolutionizing factory automation. However, robots alone are only capable of following pre-programmed commands. Robot vision is a concept that allows robots to react to changes in their task. Pairing your robots with industrial cameras gives your robots more flexibility in the face of a variable environment. This article will help you understand more about robot vision through the following topics:

  • Applications for Robot Vision
  • Industries that Leverage Robot Vision
  • When Should You Use Robot Vision?
  • Limitations of Robot Vision
  • Robot Vision Costs

Applications for Robot Vision

Robots are made more versatile with the addition of vision systems for many tasks. Without vision systems, much data is left behind. Visual information allows robots to make additional decisions that they otherwise couldn’t. Tasks that commonly deploy what is also called machine vision include:

Vision systems are helpful for many application-agnostic subtasks. These subtasks are often smaller steps that make up an overall robot application. For example, a packaging task might be made of up of smaller subtasks. Examples of these might include actions such as:

  • Pick and place
  • Measurement
  • Counting
  • Positioning
  • Localization
  • Indexing

A vision system could localize parts randomly dispersed on a conveyor belt. With this extra information, the robot can pick the parts. It can do this even though the parts are not arriving in a predictable and repeatable way. This is the power of machine vision.

Industries that Leverage Robot Vision

industries using robot vision


Robot vision is useful across a variety of industries. This is mostly due to the fact that these vision systems are so versatile. Different models can identify objects, perform measurements, inspect parts, and much more. Industries that commonly deploy robot vision systems include:

  • Aerospace
  • Pharmaceuticals 
  • Retail manufacturing
  • Electronics
  • Food and beverage
  • Metals
  • Paper
  • Plastics

Are you unsure whether robot vision is right for your application? Our network of independent advisors are ready to help! They can aid you in your automation journey from finding potential projects to receiving feedback from suppliers. 

When Should You Use Robot Vision?

It should be clear that there are additional features you gain by “giving your robot eyes”. But specific problems can be solved by adding machine vision. It’s important for you to be able to identify the kinds of issues where robot vision can help. Key features of robot vision applications include:

  • Inspection
  • Unpredictability
  • Barcode scanning or OCR (Optical Character Recognition)

Let’s look at these topics in more detail.

Inspection

It’s very common for inspection to be part of manufacturing. Inspection tasks can be found all throughout the manufacturing process. Examples of what is meant by inspection include:

  • Quality Control
  • Pattern recognition
  • Measurement
  • Counting

Manufacturers lose plenty of revenue through poor and unreliable inspection processes. Errors made here lead to part failures and rejects. Deploying robot vision inspection can catch these mistakes. What could it look like? 

Metal casting companies often utilize robot vision systems for inspection tasks. Casted metals often have irregularities that must be ground down for a smooth surface. This is often performed by large robots equipped with a grinding wheel. The size and shape of these imperfections are unpredictable. Manufacturers utilize robot vision to inspect the part between grindings. The camera inspects the part using pattern recognition looking for the proper shape within certain tolerances. The robot will continue grinding until the part falls within these tolerances.

Robot vision might not be a good choice for applications where quality control isn’t strictly defined. It might be useless to include a vision system to perform an end-of-line inspection for parts with loose tolerances. An example of this is found in foam manufacturing. A vision system is used to inspect the foam slabs for dimensions. However, foam is generally sold by weight. There might be little to be gained by performing an inspection on these parts even if imperfections could be corrected through this inspection process. 

Unpredictability

One of the most common reasons to use machine vision is to counter unpredictability. Humans can easily do this because we have eyes. If we find a part out of place, we can reach for it and pick it up. We see its location and orientation and can move accordingly to grab it. Without machine vision, robots can’t react this way. 

A common example of this unpredictability in action is bin picking. A robot is tasked with picking bolts out of a bin and placing them on a conveyor for further processing. It’s likely impossible to ensure these parts are in the same position and orientation every time. This problem isn’t so impactful with the use of robot vision. The camera can localize and orient the parts in space. This data is sent back to the robot controller. Now the robot can move to the commanded position and proper orientation to pick up the part. There are some models of industrial cameras that are capable of 3D vision. This allows them to pick through layers of parts.

It goes without saying that vision cameras won’t be helpful if you can guarantee consistent part placement. If a part is presented to the robot on a jig or any other repeatable process, you don’t gain anything by adding a camera for localization or orientation purposes. In these cases, the robot is generally able to perform its task “blind” just as well as it could with a vision system.

Barcode Scanning or OCR (Optical Character Recognition)

Plenty of manufactured materials include barcodes and/or characters on the parts themselves or their packaging. These identifiers can be used during the manufacturing process to dictate certain operations. For example, a robot vision system could be used to properly palletize two sets of packages. These packages belong to either group A or group B. Both groups travel down a shared conveyor belt. Due to the inconsistent manufacturing process, the boxes don’t arrive in any particular pattern. The palletizing robot can be paired with a vision system that can scan these boxes to identify which group they belong to. This data is sent back to the robot. With this information, the robot can place the box on the correct pallet. 

Ready to start connecting with suppliers who could help you with your project? Our proposal tool will connect you with the right companies, free of charge. Provide details about your business and what you are looking for to start receiving quotes from qualified suppliers. It only takes about 15 minutes, and you can get offers from suppliers around the world. Then just choose the best one.

Limitations of Robot Vision

Like any powerful tool, robotic vision has its limitations. Your ability to identify these limitations will help you protect your investment and avoid common issues with robot vision. 

The most common factor that can limit the performance of your vision system is light. Poor or inconsistent lighting can affect your system’s ability to correctly identify points of interest. It is important to remember that a robot vision system operates programmatically. It has set limits and thresholds that allow it to identify objects in its field of view. Bad lighting can distort how these objects appear to the camera. This makes it less reliable in the field. Supplemental lighting is often used to rectify these issues. 

Another limiting factor is resolution and field of view. Vision systems are very powerful tools. However, not every camera is as precise as the human eye. Economical camera configurations have less resolution. This ultimately means the picture is less crisp. The vision system will have less data to parse to identify smaller details. Field of view limits the size of the picture the camera is able to see. There is often a trade-off between camera resolution and field of view. Getting closer to the part can make up for poor resolution, but this can hurt the field of view. Balancing camera resolution with lens types leads to an optimal solution to this problem. 

Additionally, vision systems can’t make creative decisions. If a human operator sees something out of the ordinary, they can quickly identify this issue and take action. This might be a quality control issue like an incorrect coat of paint on a part. Perhaps the vision system wasn’t trained on color correction. While the human would know that this part has a defect, the vision system would likely let this part pass inspection. A vision camera can only make decisions based on its preprogrammed logic. You shouldn’t expect your robotic vision system to behave as you might when faced with a new situation that it hasn’t been trained for. 

Robot Vision Costs

robot vision costs


Robot vision systems are sophisticated pieces of automation technology. Their capabilities can add immense value to your automation projects. So how much can you expect to spend on a robot vision system? Like most things, you find a vast range of costs for vision systems depending on your performance and feature needs. Basic economy vision systems start around $1,000 USD. This package typically includes the camera, cables, and power supply. There are high-performance systems that can surpass $50,000 USD. These systems are high-speed and high-resolution. They are often capable of performing computationally heavy tasks such as shape recognition and machine learning. These systems generally have standalone vision controllers separate from the camera(s). 

How can you tell if your investment in robot vision is worth it? Our investment calculator will help you determine what costs are inherent in your process and what potential value can be found in improving your application. Furthermore, you will get a structured report that you can present to all the stakeholders.