Merchant Machine Vision System Integrators

The term ‘‘system integrator’‘ is used widely throughout manufacturing industries. In the machine vision market virtually all vision engine component suppliers have distanced themselves from providing turnkey systems to their customers. For the most part, some degree of customization is required to integrate a machine vision system into a production line. Hence, there is engineering content, which is often unique to the application. A vision engine component supplier would rather invest his engineering in developing new products or enhancing existing products that will yield a wide customer base for the product. System integrators would be included in that customer base.

For the most part a merchant system integrator undertakes one-of-a-kind projects. In some cases, their customer may actually equip more than one line with the same system, but the system has been configured for that one customer’s unique requirements. In the case of system integration of machine vision systems an integrator’s primary skills may actually be in a discipline other than machine vision: material handling, assembly machines, bar codes, robotics, packaging lines, controls, etc. In these cases, machine vision becomes one of several complementary sensor technologies that they integrate as part of a project mostly related to their primary skills (material handling, packaging lines, controls, etc.).

Their value adder relates to the integration of all the components for the specific project undertaken where machine vision is only one of those components. In addition to integration into an existing line or plant, the value added might include the development of a graphic user interface reflecting the application of the turnkey solution developed for the customer. In the case of machine vision, the value added may include some application engineering associated with the ‘‘staging’‘ – configuration of the lighting, optics/camera, part and their brackets. More often than not the staging design skills of these type integrators is limited to the class of applications for which they integrate machine vision – bar code/2D symbol reading, packaging lines, etc.

Significantly, there are integrators of machine vision systems where their primary business is just that – the integration of machine vision systems. These integrators include multidisciplinary skills on their staff – skills in optics, lighting design, cameras, etc. These integrators are much more than ‘‘point-and-shoot’‘ integrators of machine vision. They will undertake projects that entail significant application engineering – multiple camera/multiple lighting arrangements, software development, etc. that are often required to satisfy unique application requirements. While there are many ‘‘point-and-shoot’‘ integrators of machine vision systems, there are not that many whose skills are first and foremost machine vision integration. Many of these can be found listed in the AIA Directory of machine vision vendors.

Input for this article on the challenges of integrating machine vision was canvassed from the companies listed in the AIA Directory. The following chose to respond with substantive insights into what is needed to succeed in a machine vision project.

Perry Cornelius – ABCO Automation
Brian Durand – BD Automation
Brian Smithgall – Image Labs International

Can you briefly describe the services you provide as an independent merchant system integrator of machine vision systems? What other value adding services do you provide? What industries do you serve? 

Perry Cornelius (ABCO Automation): ABCO Automation provides turnkey integration of robotics, machine vision, and custom automation.  For machine vision, ABCO provides evaluation of customer’s samples in our vision lab.  Detailed documentation and training are provided with our systems; detailed system validation is also available.  ABCO serves mostly the automotive, consumer goods, and medical device industries.

Brian Durand (BD Automation): BD Automation designs and builds vision machines. These turnkey solutions use machine vision to inspect a product, or guide a robotic operation. Services include mechanical, electrical and optical design, machine vision and control system programming, fabrication and assembly, factory testing and validation, startup, operator training, and extended support. Though we have systems installed in many industries, most of our customers manufacture medical devices.

Brian Smithgall (Image Labs International): Image Labs International develops machine vision systems for end user applications. We work with customers on initial design concepts, determining specifications, detailed design, implementation, and testing. Our designs include custom material handling, lighting, and optics, as required by the project. Our engineers work with all the major system component providers, so we stay current on the latest technologies. We also enjoy working with smaller customers who need help selecting equipment for their own projects.

What type of machine vision hardware do you normally integrate? Frame grabbers, smart cameras, embedded vision processors, etc.?  

Brian D.: We have the freedom and expertise to select the platform best suited to the project, be it a smart camera, embedded processor, or PC based system. We find that smart cameras have advantages in simpler applications, but many medical device applications benefit from the greater power and ease of use associated with PC solutions.

Brian S.: Since 1993, the company has worked with all levels of imaging systems, from board cameras to multi-spectral imagers. We have implemented factory solutions with frame grabbers, smart cameras, and to a lesser extent embedded vision processors. While the smart cameras are great fits in certain applications, PC-centric solutions are still most common.

Perry:  Smart cameras are sufficient for many of ABCO’s machine vision projects.  However, more challenging applications (e.g., linescan, ultraviolet, infrared, very high speed) still require advanced, dedicated systems such as frame grabbers with application specific cameras.

Has integrating machine vision become easier or harder over the years? Why? How has this impacted the price a customer pays? 

Brian S.: It is easier in the sense that more powerful equipment is available, but as abilities rise, so does the complexity of the customer’s expectations. Projects that were previously too difficult are now feasible. With the lower cost of some hardware, development costs have come down slightly, but recurring costs have dropped significantly.

Brian D.: Of course the answer is ‘‘both’‘. Integration has become easier in that the hardware is more robust and the software algorithms more forgiving. There are also more component options than ever before. The challenge is to know when to use which component, which is best learned from experience. Customer expectations continue to rise, challenging us to innovate and always improve. While the cost of integration has decreased for many common applications, one continues to see demanding projects that require sophisticated hardware and advanced expertise for best results.

Perry: Easier. Faster vision processors, better software, and a wider variety of (and better quality) lighting products.  The overall price is definitely lower than in years past.

Is the value added by merchant machine vision system integrators greater or less than it was five years ago? Why? 

Perry: For us, I would say about the same.  Customers still depend on us to recommend the proper lighting, optics, and imaging algorithms to provide a reliable & robust system. 

Brian D.: The amount of value added by machine vision system integrators is dependent on the project. Frankly, we have little value to add to a $2000 vision system hung above a conveyor that a distributor sales person is willing to install for free. On the other hand there are projects where customers cannot afford to fail, cannot afford to have false rejects (or worse, false accepts), and cannot afford recurring downtime. As with all automation there are instances where it makes sense to try a cheap fix, and there are situations where hiring the expertise and hard-won experience of an integrator is money very well spent.

Brian S.: With the availability of more powerful tools, integrators can include more powerful features like databases, SPC, or robotics. In a sense, customers are still coming to us because the projects are beyond their interest or expertise, yet they understand the job can be done by a specialist.

What are the challenges in integrating machine vision these days? 

Brian D.: The challenge is no longer to get machine vision to work -- the new challenge is to apply it well. Developing the most reliable and easiest to use solution, and delivering it on time, is what keeps us on our toes.

Brian S.: There are rapidly changing technologies, which mean continued risks of compatibility and robustness. It is the integrators job to steer clear of these pitfalls.

Perry: Technically, one challenge is dealing with Ethernet type communication issues with other controllers (e.g., PCs, PLCs, robots …) which are reminiscent of RS-232 type issues from years ago.  Commercially, it is sometimes challenging to justify the cost of the value added an integrator.  Customers see ads for a vision system for a few thousand dollars and are then shocked by the price to integrate it.

What is the role of a customer in a machine vision integration project? 

Brian S.: Ideally the most important role of the customer is to clearly define the scope and requirements of the project.

Perry: The roll should be to provide detailed specifications for their application, scheduling of adequate downtime for installation, and arranging the proper personnel to be adequately trained.  They need to ensure their personnel will take ownership of the system.

Brian D.: The most important thing a customer can do is to communicate their needs prior to awarding a contract. Its critical the integrator understand all flaws that need to be detected, as well as normal production variances that are acceptable.

What is the role of the vision hardware supplier in a machine vision integration project?

Perry: For us, the vision hardware supplier is used mostly to provide technical support for their hardware (i.e., handling of bug issues and detailed feature explanations).  We seldom use the hardware supplier to evaluate an application unless there are cycle time issues or the performance of a feature needs to be verified.

Brian D.: We believe the role of hardware suppliers is to support the integrator when needed.

Brian S.: Equipment providers must provide reliable building blocks, with sufficient documentation and support to build applications.

What is the role of the system integrator? 

Brian D.: The role of the system integrator should be to accept complete responsibility for delivering a quality solution, as specified by the customer. The best integrators have experience with a variety of vision platforms, and will suggest which components will yield optimal results. A good integrator will never try to pass blame onto one of their vendors. Since many applications are dependent on good presentation of parts to a camera, the best integrators will also offer internal mechanical design and fabrication services.

Brian S.: The integrator must have access to a variety of tools that he knows are proven and reliable. They must understand the project development process, which will include the customer in the process. They must communicate clearly at all stages in the project, and ultimately deliver a system that works and works well and that they will support.

Perry: Initially, a thorough evaluation of samples should be performed in order to provide the customer with a realistic idea of what the system will be capable of inspecting.  After that, the integrator is responsible for providing a system as promised (e.g., design, programming, detailed documentation, training, validation, etc.).  If there are issues along the way, the integrator should immediately inform the customer (i.e., effective project management).  

What are some mistakes buyers of machine vision systems make? 

Brian S.: There are two opposite extremes here. First, a naïve buyer will assume the system will do everything and can adapt as easily as the human to visual situations. Secondly are buyers who have seen others burned by the first category, and are decidedly against any future vision projects. If a project is clearly defined and well executed by the participants, it should succeed like any other engineering project.

Perry: A major mistake we see is an end-user (or distributor) attempting a project that is beyond their expertise.  The end result is normally a system that is non-reliable and that needs to be frequently adjusted.  Machine vision has come a long way, but many are still learning the hard way.

Brian D.: Everyone is in a hurry these days. There is the temptation to assume that all integration proposals will yield similar results, and purchase decisions are, therefore, made based on initial price. The irony is that may projects are justified based on reducing defective product escapes and reducing labor. We have seen many cheap inspection systems that pass too many defective products, and require too much labor to keep them running. Often times spending a little more will provide a better return on investment, but only experienced customers know this.

What advice would you give a prospective customer for a machine vision system integration project? 

Perry: If they are unsure of the application, the guidance of an experienced, reputable integrator should be considered.

Brian D.: Before placing any purchase order we suggest documenting your needs and priorities, giving the integrator as much sample product for evaluation as they will tolerate, and then getting to know the integrator. Visit their facility. Meet their engineers. See their vision lab. Learn about similar projects they have completed. The best integrator for the project will become readily apparent, and the project will be a success.

Brian S.: Find someone with experience. Experience is more than knowledge of the tools, it includes experience in vision applications like specialized lighting and also successful project development and implementations. An integrator that can provide a formal engineering development process, with formal specifications, design reviews, testing, change management, documentation and support structure is valuable. It is important to get a demonstration of their ability with your product or a similar one, even if you have to pay for it. It helps prove the technology and test drives the company relationships.

What changes in machine vision do you anticipate in the next 2 – 3 years? Will these make things easier for a machine vision system integrator? How will this affect the ultimate price to a customer? 

Brian S.: Things will keep getting faster and smaller and more intelligent, which will open up more opportunities. For a specific application, this will mean simpler integrations and lower prices.

Perry: Continued price reductions with performance increases.  Applications not considered previously due to cost & feasibility will be considered (e.g., high resolution, infrared, color).   Cost justification of using an integrator may be affected as hardware prices decrease. Therefore, integrators may be pressured to decrease pricing, thus lowering the ultimate price to the customer.