- THE MAGAZINE
- WEB EXCLUSIVES
About a year ago, I wrote a series of columns on open source software and Linux for use in machine vision, sensors and factory automation applications. I’d like to use this month’s column to follow up on some of those ideas and give some insight into where things could be going.
In machine vision and factory automation, we often follow, or are carried along by, trends in much larger and higher-visibility industries. Machine vision often is carried along by the security market, for example, while sensor systems and factory automation have made great strides taking advantage of traditional PC and IT technologies such as Windows, USB, Ethernet and WiFi.
This makes me wonder what current trends might be out there that could allow us to predict the future of machine vision and sensor system deployment. Here are my five major candidates.<
LinuxIt is here to stay. My company has been using it for most new major projects for more than a year and the advantages are compelling. Support for more and more devices is available, and new hardware and software products are much more likely to be Linux-ready than they were a year ago. The cost of Linux and all of the support services that one might need to make it into a complete system is truly zero. Absolutely zero. Junior programmers are not only familiar with Linux, but in many cases they prefer it over the old, traditional Windows operating system (OS). So as Linux grows, it will be more apparent in vision and sensors.
AppleNow the world’s largest technology company, more than just back from the brink of extinction, Apple is dominating just about every market it enters. Supplying both software and hardware affords huge advantages, a fact IBM was once able to capitalize on before the PC revolution.
While Apple can’t be argued to be much of a promoter of open systems, the Mac OS X is a Linux-based system, so companies migrating to Linux software are inherently more Mac-like than Windows-like. This is a stealth win for Apple, and may not be as well-known as it should be.
Smart DevicesThe proliferation of GUI interfaces, not just on PCs and Macs, but on the ever-more-popular smart phone, has made the graphical interface required, not just optional. Command lines are fine for engineers and programmers, but users want snazzy graphical interfaces that are easy-to-understand, simple and robust. Customers want an iPhone or Blackberry app that allows them to adjust a sensor threshold or focus a camera on the factory floor. We can do this so it will become commonplace. My next megatrend provides a clue about how.
QtQt, a software system from Norwegian Trolltech, now owned by Nokia, is a software development framework embraced by programmers of smart phones, PCs, Macs and Linux workstations. You may not have heard of it, but programmers are very familiar with it.
It provides a programming model that allows a company to deploy software on multiple GUI platforms without redesigning and recoding every time. The Qt (theoretically pronounced “cute”) software world provides design tools that allow incredibly rich and beautiful graphical interfaces. Design a program that runs on Windows-a few tweaks and a recompile, and you have a Linux, iPhone or Android version. Many current applications which simply make use of the graphical interface portions of LabView or WonderWare can easily be programmed in Qt, and then redeployed on Macs or smart phones, Linux systems or PCs, with nothing more than a recompile.
This makes multiplatform software design much easier, even for smaller software houses, and undoubtedly gives us a glimpse of a future in which the underlying OS will not be a choice that locks software to an OS as has happened so much with Windows.
My company is currently developing a series of applications for factory automation and machine vision using Ubuntu Linux. I have been incredibly impressed with the stability and broad feature set available at absolutely no cost for Ubuntu. A free KVM virtualization system allows Windows to be installed as a virtual machine on the computer if there are specific Windows-only tasks that must still be handled.
Our GigE vision cameras, motion control systems, illumination systems and I/O cards all function as well or better than they did in Windows, and our ability to stream images to disk for later post-processing is dramatically better than anything we could expect from Windows.
As a small company, we can usually afford to be out at the forefront of new component deployment since we are often developing highly technical and unusual solutions from scratch. It is much easier to make a leap when there is not a huge embedded legacy software base that needs to cross the chasm with you. But based on our experience converting Windows applications and old DLL libraries, the job is probably easier than you think, and opens a great way for end users, OEMs, integrators and consultants to get the job done with lower cost hardware, lower software licensing fees and less proprietary technology than ever.
And accomplishing that is my fifth and final megatrend.