• Font Size:
  • A
  • A
  • A

Feature Articles

How to Help Customers Maximize Their Machine Vision ROI

by Winn Hardin, Contributing Editor - AIA

The machine vision industry faces a Goldilocks complex when it comes to applying the appropriate type and amount of technology: Make the system too simple, and it might miss important inspection parameters. Make the system too complex, and users are paying for products and services they will likely never use. How do you create a vision system that will land in the middle and provide customers a return on their investment? It’s a big question that vision OEMs and integrators face out of the gate, and they answer it by investigating every application detail to design a system that fits just right.

A proof-of-concept or feasibility study is the first step in measuring ROI, since it lays out exactly what the customer can expect from their vision system. However, this study also represents the first step in investing in the system, as many integrators charge for this service. “You might pay $10,000 for a proof of concept, but it could save you $100,000 down the road for it to function properly,” says Sean Lett, Vision and Automation Sales Manager for machine vision integrator Radix, a division of AIS Technologies Group. “The companies that are truly engaged in improving their process and controlling product quality will invest with you.”

From that study, the integrator creates a custom-engineered design that “focuses on what the system is trying to accomplish and how it will verify what it accomplishes,” says Tom Brennan, President of machine vision integrator Artemis Vision. “That could include testing, validating, and proving out that the software is detecting the desired outcome, or running specific challenge parts, calibration parts, or other reference materials to establish that the system is functioning as designed.”

This process relies on clear communication between the integrator and customer. Some manufacturers believe that a camera installed above a product or part will be enough to perform the necessary inspection. While a simple application may need only a smart camera, projects that involve system integrators typically require more complex designs.

“We spend a lot of time, especially with new customers, working on functional requirements and walking them through the machine vision process,” Lett says. “With our final documentation, the customer knows exactly what they’re getting.”

Customers who take an active role in the vision system development process have better outcomes. Take the example of inspecting labels on water bottles. Some manufacturers still rely on manual inspection: an operator checks a sample bottle every 20 minutes to ensure that the labeling is correct. It is an application that would benefit from machine vision, and the integrator’s responsibility is to show customers why.

“If you missed labels on 1,000 bottles, the cost per bottle for water isn’t that great to help justify ROI of a $30,000 vision system,” Lett says. “But if you are doing energy drinks or alcoholic beverages, you’re not talking pennies lost in the bottle inspection. You’re talking dollars. Those 1,000 bottles justify the ROI.” Typical ROI parameters include reducing labor costs, catching errors quickly, eliminating the cost of a rework, and even protecting the brand from distributing faulty products.

Getting What You Pay For

Cost reduction continues to be the primary parameter that machine vision customers want to measure. Artemis Vision recently sold a system that saved one company hundreds of thousands of dollars upfront via a simpler design. The customer needed to measure the dimensions of foam sheets used for General Electric wind turbine blades. A laser profiler is often recommended for this type of job because it can take perfect dimensions in all three axes.

“After we talked with the customer, it was clear they didn’t want that $500,000 3D laser profiling-based system,” says Brennan. But they sought something better than their current process, which was a worker using a tape measure. Artemis Vision created a simple solution: an overhead camera installed above a conveyor to measure length, width, and squareness of the material and single-point lasers to determine thickness.

“Our customer won the contract by showing GE that when you place orders with them, you are going to get consistent products made to specification,” Brennan adds.

In another example of using just the right amount of vision, hardware and software manufacturer Matrox Imaging has modularized its software offerings. Users can select runtime licenses from a menu of runtime modules, activating only the licenses for the modules they need. “That way they aren’t paying for anything they don’t actually want or use,” says Sam Lopez, Director of Sales and Marketing for Matrox Imaging.

Once a customer has developed their application, they run a utility included in Matrox Imaging Library (MIL) software that tells them which functions are used in their application and therefore which modules should be licensed from Matrox Imaging. Tools built into the software — which can be deployed on any hardware platform — allow users to run the software’s benchmarks so they can quickly see whether they are within the performance goals needed for the application to work.

Measuring the ROI of off-the-shelf products is typically more straightforward than doing so for their fully engineered counterparts. Radix’s Tool Tracker vision-based system tracks a torque gun or other handheld tool to verify that an operator takes corrective action on an engine’s loose bolts. For one customer, Tool Tracker eliminated the need for the operator to go through every single bolt on a specific part of the engine. The product brought a 10-minute process down to two minutes, saving the customer $96,000 a year in processing and time.

More and more, manufacturers are using data generated by their vision systems to improve ROI in other areas by performing analytics on their processes. “Internal to their process, we’ve seen customers look at whether supplier material correlates with defects, and then pushing back on suppliers,” Brennan says. “We’ve also seen them trying out new saw blades for a cutting operation and running the machine at a slightly higher rate while using the data from the vision system to see if the process is drifting out of control.”

External to their process, one customer uses data for all the grades of product it makes to drive better pricing. “They know what percentages of product meet what specs and can upcharge customers who demand tighter tolerances,” Brennan says.

Roadblocks to ROI

While it’s important for OEMs and system integrators to perform due diligence to maximize a customer’s ROI, they may run the risk of taking too long to develop a vision system. “People often overlook the time it takes to actually develop an application,” Lopez says. “It’s all well and good to deliver a great solution to a problem, but if it took you two years to get there, it becomes very expensive. It’s in our interest — as well as the customer’s — to get up to production speed as quickly as possible.”

To that end, Matrox Imaging developed agile prototyping and development tools, such as MIL Copilot. MIL Copilot is an assistant for creating vision systems that allow programmers to investigate, design, verify, and generate code for their application. “You’re saving a lot of time and expense for the customer because they don’t have to write their own code from scratch,” Lopez says.

Vision OEMs and integrators have many ways to measure a customer’s ROI, but it all comes back to how the system is developed in the first place. When establishing ROI parameters, the integrator should not confuse confidence with arrogance — the latter can lead to overselling, overpromising, or completely missing the vision system’s goal. “It’s unhelpful to walk in the door smugly and say, ‘I’ve done this inspection before,’ when in fact every application is different,” Lett says.

Sometimes understanding the importance of ROI comes from making mistakes. Years ago, Lett sold a system that gauged on metal parts. Two days after the system was installed, the customer called to say they were getting a lot of false failures on good parts. When Lett visited the customer, he saw that the parts were covered in oil and surrounded by metal shavings.

“I looked at the customer and said, ‘That’s not the same part I tested on. I tested on nice, dry, clean parts,’ Lett recalls. The customer countered that his parts were never clean. “We are looking at each other, and I said, ‘Well, you didn’t tell me that.’ And he said, ‘You didn’t ask.’

“Learning from our mistakes makes us better integrators,” Lett says. And it puts OEMs, suppliers, and integrators in the “just right” Goldilocks zone for developing vision systems that maximize ROI.

 

 

Comments:

There are currently no comments for this article.


Leave a Comment:

All fields are required, but only your name and comment will be visible (email addresses are kept confidential). Comments are moderated and will not appear immediately. Please no link dropping, no keywords or domains as names; do not spam, and please do not advertise.

First Name: *
Last Name: *
Your Email: *
Your Comment:
Please check the box below and respond as instructed.

Search AIA:


Browse by Products:


Browse by Company Type: