Wednesday, December 5, 2007

Blue rinse group

Machine vision looks more like a bargain when it's viewed as another helping hand... It's always good to save money. And with this in mind, I decided that I would fly to October's Pack Expo show in Las Vegas, NV, using a “vacation package” from Southwest Airlines. For a cost of less than $600, the airline flew me from Manchester, NH, and back, booked me three nights in the Circus Circus hotel/casino/resort, and gave me complimentary passes to a day trip to the Hoover Dam. Upon boarding the aircraft, I discovered that “Blue Man Group” was playing at the Venetian Resort for a price of just $70. Along with the Hoover Dam trip, this was an event I would never attend.

Looking around the aircraft passengers, it appeared many of them were of the “blue rinse” group—people aged 65 and older who were on a similar low-cost trip destined for Las Vegas. Why anyone would want to go to a town reputed to have been founded by gangsters in the middle of the desert still amazes me, but I was later to find out. After interviewing several passengers on the aircraft and at the hotel I discovered the reason. It was very inexpensive. While some companies attending Pack Expo had booked their employees into $200/night hotels (some taking hundreds of their customers to Las Vegas entertainment), I was stuck in Circus Circus with the blue rinse group, where the average cost of a room is $60. On reflection, perhaps this was not one of my wisest decisions. After battling though a myriad of slot machines, I went to bed.

The next day, I decided to take a cab to Pack Expo. In more than five halls, most the size of two baseball fields, nearly every packaging company in North America displayed products and machines that produce, fill, label, wrap, and inspect bottles, paper packages, and plastic containers. With more than 2000 exhibitors and more than 20,000 attendees, the show was, like Las Vegas itself, over-the-top, outrageous, and extreme.

Although my trusty companion, Judy Leger, could not attend the trip, she had supplied me with a list of companies to visit. Without it, I would have been lost. In the space of three days, I managed to visit about 20 companies that incorporate or produce machine-vision products used in the packaging and production industries. Many of these systems can be found in this issue of Vision Systems Design. Traversing the show floor, I also had the opportunity to visit companies that simply had not embraced the concept of adding machine vision to their production equipment. While many produce machinery that manufactures, wraps, and fills specific products, very few—in fact fewer than 20—were exhibiting machines that provide any kind of inspection of the finished product. It all seemed rather odd.

On the last day of the show, I sat at our tiny booth contemplating the subject. I was then approached by an engineer who needed to perform web inspection of wrapped film traveling at 400 ft/min. He had been quoted a price by a system integrator of $20,000 for a system that could accomplish the task. He felt that the price was far too high. He wanted something for less than $15,000.

After I explained that the cost of linescan cameras, Camera Link frame grabbers, lighting, and software would probably be about $15,000, and nonrecurring engineering costs would be at least $15,000 for his application, he left our booth rather amazed. I could not help wondering, however, whether the bill for his trip to Pack Expo would be more than $2000. Perhaps those who need to expense machine-vision systems need a different model.

As Preben Hjornet, chief technology officer at InMoTx (Sacramento, CA, USA), suggested to me, “Engineering managers should perhaps regard any automated system, whether it uses a robot or other automated equipment, as an employee rather than a machine. That way, the machine would not be regarded as a capital expense.” Until that happens, however, many in the packaging industry will still regard machine vision as an expense they cannot afford.

Wednesday, October 24, 2007

Engineering in 2010

Every engineer likes block diagrams. Not only do they encapsulate the concept of a system design at a high level...Every engineer likes block diagrams. Not only do they encapsulate the concept of a system design at a high level, but they are easily understood and form the basis for the design of nearly every machine-vision system. Over the past ten years, the many authors that have contributed to Vision Systems Design have supplied these diagrams to allow you, the reader, to rapidly understand the hardware concepts behind system design.

But of course, these block diagrams can only tell part of the story of any design. The diagrams are, by their nature, static objects that only provide a conceptualized nature of any design. Engineering, on the other hand, is a dynamic process that involves interfacing numerous hardware components, overcoming timing issues, properly configuring software, perhaps on multiple processors, and then debugging the system to ensure correct operation.

In developing a machine-vision system, this task is more complex because of the nature of the parts that need to be inspected and the defects or parameters that need to be measured. What is, in fact, required is an integrated software-development environment that can be used to describe, simulate, test, and build a system in software alone.

While this may appear challenging, it will probably form the basis of all engineering design within the next ten years. Indeed, software developers are already blazing a path toward this goal with graphical development tools that allow the user to develop and test image-processing software before deployment. However, for the concept to become a reality, tests on individual parts will need to be simulated in CAD environments.

Electronic components such as cameras, lighting, frame grabbers, I/O controls, switches, and PLCs will need to be simulated as virtual objects within a graphical user interface. Timing and control data from the electrical CAD models of these products will need to be stored in databases. Host CPU operating systems and image-processing software will need to be modeled as objects that control this virtual environment.

At present, many of these ideas exist in isolation. CAD packages can be used to simulate the mechanical motion of systems. Electrical specifications for the components used may exist on data sheets. Machine-vision software must be developed on realtime or Windows-based computers in isolation from the hardware that will be used when the system is deployed.

Imagine, however, using an integrated graphical system environment to develop a machine-vision system. The supplier sends you a CAD model of an ideal part and CAD models of parts with defects. By using simulation software, a complete machine-vision system is developed as a mechanical abstraction in which lighting, illumination, and lens parameters can be controlled. These images are used as a model with which to develop machine-vision software. Simultaneously CAD and electrical models of the hardware are described in a CAD model.

At the highest level, the engineer can visualize the design as a block diagram. As he or she probes deeper into the design, the mechanical, electrical, and software models are revealed so they can be tailored for a specific application.


This, in essence, transforms the static block diagram into a dynamic software model. Engineers that endorse this concept will reap the benefits in numerous ways. By sharing such “block diagrams” with other engineers, systems will be able to be developed more rapidly without having to re-invent the wheel every time a new system needs to be developed.

These concepts are already being explored by a number of companies worldwide. When realized, the simple block diagrams of the past, published electronically, will set a new standard for system design. When published electronically, such sophisticated diagrams will also benefit the less noble art of electronic trade publishing.