When Does It Make Sense to Use a Machine Vision Consultant

Defining variables, providing valuable training, avoiding costly mistakes, and providing technical competence are only a few of the most compelling reasons for using a machine vision consultant when user companies begin defining their requirements and decide to employ machine vision in their manufacturing lines.

'Machine vision is a very specialized technology,' points out Marcel LaFlamme, vice president, RVSI Northeast Robotics (Weare, NH). 'It requires knowledge of cameras, optics, illumination, and processors and there aren't too many who have a thorough knowledge of these four technologies. Most of the people involved with the industry have mechanical and robotic type experiences, but very few have the other four.'

Timothy P. White, founder and president of Northeast Robotics (New Boston, NH), reinforces this viewpoint. 'If a machine vision project requires expert guidance and there is no one in-house with broad experience, it is valuable to have a consultant who has knowledge of lighting, cameras, software, and machinery,' he advises. 'Generally, the cost of installing machine vision is small compared to the cost of not doing it right. And, generally, the major cost can occur after the system is installed, trying to make it work. So, the efficiency one gains by hiring a consultant is to vastly improve the probability of success from day one.

'The work often begins after the vision system is installed,' continues White. 'How is the vision system going to be managed after it is installed? Who's going to manage it? Who's going to be trained to run it? Who's going to tweak it? Who's going to be trained to maintain it? Machine vision is not something you put in and never touch again. The shop floor personnel should be able to handle it. Installing a system which then requires an expert to maintain it is guaranteed failure and heartache.'

Of course, there are a number of other cogent reasons to use a vision consultant, not the least of which is experience and knowledge of available solutions, particularly in the lighting field. 'Lighting has changed vastly in the last 10 years,' continues Tim White. 'There's been a tremendous evolution in technical lighting, particularly with regard to the imaging of specular objects-foil, glass, and polished metal. Inspecting such parts 10 years ago was impossible. There's a tremendous repertoire of illumination solutions specific to specular imaging that did not exist 10 years ago.'

In other instances, a consultant's breadth of knowledge can be often used to identify how users should organize what the issues are and help them jump-start some of their initial projects. 'It really boils down to two things,' explains Perry West, president, Automated Vision Systems Inc. (Los Gatos, CA). 'In one case, the consultant has specific expertise that can help the user company define its objectives, and, in the second case, where the user company has already allocated all its resources and it needs to expand its resource base.'

In one user example, Sataki USA Inc. (Modesto, CA), one of the world's largest producers of rice milling equipment used in rice milling, husking, and screening, was able to move into the field of food processing by utilizing the services of Perry West. In this case, West was able to advise on the optical system used in the length grading of almond nuts. 'He designed and constructed the optical system along with the software development for recognition and measurement in the ejection control of a machine that was equipped with a belt sorter used as conveyance for the almond nuts,' explains Jerry W. Brum, manufacturing engineer at Sataki USA. West researched the optimum optical means for the system, selected the cameras, computer technology, and also developed the software for the system.

'As a result of West's involvement, we were able to keep our internal engineering dedicated towards larger projects, while we were able to gain a return on our investment that we normally would not have received in a market that we normally wouldn't get into,' sums up Brum. 'From a monetary standpoint, there was a big return on our investment.'

Apart from the return on investment, user companies can often avoid making costly mistakes. And those mistakes can be both financial and political, as pointed out by Ken White, president, Visual Sence Systems (Ithaca, NY). 'If my work doesn't pay out at least one to one if not ten-fold, then I'm doing something wrong,' he opines. 'The financial return is one but the political is also important. Because to champion a path through a major portion of a project life cycle, and then to have to change gears because you missed something, is politically very damaging.'

In one case, White points out, a company that manufactures counter tops and floor covers invested in a spectro radiometer, which measured average color in order to make sure that their products were labeled correctly. The products were similar. The human eye could easily differentiate between two very similar products.

'The company spent nearly $1 million on the radiometer, only to find that it was technically impossible,' says White. 'In order to differentiate between the two, the equipment had no sensitivity to texture and often came up with a wrong number. So, the equipment made mistakes on a high percentage of the time. The political cost to use the right technology in this application was incredible. And the financial cost was also very high. If they had used the right technology, it would have cost them $20,000 per line, instead of the $80,000 it cost them. This is an example where they wasted $60,000 per manufacturing line. And this company had several lines,'

Vision consultants are valuable not only because of the breadth of their knowledge, but also because they can gauge the technical feasibility of a proposed application. 'One of the many things we can do for users is to define their requirements into a comprehensive functional specification,' explains Nello Zuech, president, Vision Systems International Inc. (Yardley, PA). 'We've done a lot of these specifications, but they're not equipment specifications. A systems integrator would be more likely to write an equipment specification rather than a functional specification. People often don't understand the difference between these two types of specifications. The issues here are the variables that users have to contend with. Those have to be defined. They could be variables dealing with appearance, position, environment, components, and all the algorithms that might work under one set of circumstances and might not work in another. It is critical that those variables be defined.

'The next part of the project is to define the training, write the specification, and provide users with a list of qualified vendors,' continues Zuech. 'This means that not only do you provide a list of vendors from your database but also you would contact the vendors, describe the application, and make sure that they do have the relevant experience and the capability to do the job. Today, people in the vendor community are becoming more selective. You have to go through a kind of qualification before preparing a bid package and turning it over to the client.'

The knowledge, experience, and know-how of consultants are all critically important to user companies in defining their applications. Consider the experience of Smith + Nephew Inc. (Memphis, TN), a manufacturer of orthopedic implants for the medical industry. By choosing the right type of vision system, the company was able to avoid its biggest fear-the cost of product recalls both financially and in terms of reputation and image.

The problem the company had was to ensure that right and left orthopedic implants were correctly labeled, packaged, and shipped. Nello Zuech provided the company with a complete report based on his findings of the manufacturing line. 'We chose a PC-based system rather than a dedicated processor,' says Ron Stephens, manufacturing engineer at Smith + Nephew. 'Zuech was able to confirm that the company was going down the right path in its choice of a vision system.'

The problem that the company had to solve was one of correct packaging and labeling of its stainless steel and titanium parts. 'We had problems in the past, where the vision system would identify the correct part in the box, but, by the time it was put into the box and the label was put on the box, we occasionally had a part that was put in the wrong box,' recalls Stephens. 'We brainstormed the problem, because we wanted a 100% guarantee that this would never happen.

'We came up with a solution that combined an X-ray system that would see the part inside the box and a vision camera that would read the outside of the box and would also read the bar-code label on the outside of the box,' continues Stephens. 'And, if that bar-code matched the physical shape of the item inside the box, we now knew 100% that the part inside the box was the right part and the label on the outside was correct.'

Stephens is of the opinion that money spent on the company's vision system is money well-spent, because the cost of a recall involving hundreds of distributors all the way down the supply chain would not only be financially ruinous, but also' very bad for business.'

James F. Manji is a free-lance writer in Brunswick, OH.

.