- THE MAGAZINE
- WEB EXCLUSIVES
Monitoring temperature is essential to several machine vision applications. Knowing that a temperature is constant or remains within a safe range ensures quality. Being alerted in real-time to a temperature flux necessitates immediate action to maintain safety and efficient operation. Go/no-go; repair at next scheduled maintenance; or shutdown immediately? Making the right decision quickly will directly influence the status of your equipment, personnel, production and bottom line. Relying on infrared (IR) imaging takes much of the uncertainty out of those pivotal decision-making moments.
This necessitates immediate action to maintain safety and efficient operation. Go/no-go; repair at next scheduled maintenance; or shutdown immediately? Making the right decision quickly will directly influence the status of your equipment, personnel, production and bottom line. Relying on infrared (IR) imaging takes much of the uncertainty out of those pivotal decision-making moments.
In this article, we’ll cover some, but certainly not all, machine vision applications of IR imaging. We’ll then recommend some questions to ask a vendor prior to implementing IR into your machine vision monitoring. It’s important to note, however, that machine vision applications for IR imaging are being invented daily. As the technology improves and becomes more affordable, more people are getting exposed to its possibilities. As exposure increases, so does the creativity.
An Infrared OverviewSimply put, IR imaging involves pointing an infrared camera at a particular scene where you need to see more than what you can with the naked eye or through a visible light camera. Infrared covers the portion of the electromagnetic spectrum from approximately 900 to 14,000 nanometers (humans can see in the 400 to 700 nm range).
IR is emitted by all objects at temperatures above absolute zero, and the amount of radiation increases with temperature. A properly calibrated IR camera can capture thermographic images of target objects and can provide accurate noncontact temperature measurements of those objects.
For brevity’s sake, I’ve categorized three types of machine vision IR imaging: on-the-move, stationary and proxy.
On-the move: IR camera is capturing images and data of moving targets, such as on a conveyor belt.
Stationary: targets are stopped and examined as part of a process where the IR camera is the go/no-go decision maker.
Proxy: IR camera thermographic output is being used to replace a mechanical interruption of the production process.
These definitions will become clearer through the following use-case scenarios.
On-the-Move ApplicationsMoving targets are involved in many different applications, from plywood mill machine monitoring to food production to coal transport.
In the plywood milling application, steam from an open vat of hot water obscures the machinery operator’s view of the logs as they are maneuvered for proper alignment in the log vat. An IR camera can present an image to the operator that makes the cloud of steam virtually transparent, thereby allowing logs to be properly aligned in the log vat.
At a food processing plant, the line operator can use an IR camera to locate undercooked items within randomly scattered parts and then remove them from the conveyor.
And in a coal plant, coal could spontaneously combust anywhere along kilometers of conveyor belt during transport from piles to boilers, potentially causing disastrous results. By installing an IR camera or cameras that can view coal passing on the conveyor belt, operators can see coal that’s on the verge of combustion.
Stationary ApplicationsInfrared imaging can be used when IC testing on circuit boards. Heat loading on integrated circuits cannot be adequately characterized during overload testing. The solution is that an IR camera captures images during and after a current pulse that lasts only 800 milliseconds (ms) to characterize temperature rise and fall. With a 60 Hertz frame rate, a new frame can be captured about every 17 ms (50 frames/800 ms).
Commercial vehicle brake examination at a vehicle weigh station is another potential application. Faulty brakes, bearings and tires in commercial vehicles are more likely to result in a vehicle accident. The solution? By mounting an IR camera in the entrance of a weigh station, detailed thermal images of the undercarriage of commercial vehicles can be captured as they enter the station. Computerized comparison of the signal patterns with predefined conditions will detect anomalies.
Testing car seat heaters during production can also benefit from infrared imaging. Using contact temperature sensors to assure proper operation of optional car seat heaters slows down production and is inaccurate if sensors are not properly placed. The solution was to have an IR camera detect thermal radiation from the heater elements inside the seats and provide an accurate noncontact temperature measurement.
Proxy ApplicationsPressure casting mold in automotive manufacturing requires temperature monitoring. Using an X-ray system to reveal subsurface imperfections on a safety-critical part caused by improper temperature distribution of the part’s mold resulted in high scrap rates. X-ray exams couldn’t be done inline and took a few hours to reveal imperfections, so all parts molded during that interval had to be scrapped.
By inserting inline temperature measurement of the part mold as a proxy for the X-rays of the actual part, an IR camera system enabled the mold operator to check and adjust the temperature distribution of the mold in real time. The interval between X-ray capture and analysis results was eliminated, hence, so were the high scrap rates.
Six Questions on IR ImagingIf you already have visible light cameras in your system, then you’re already familiar with some of the factors to consider when integrating IR imaging cameras (such as durability, field of view and remote control). Just in case, here are six questions you need to ask of any prospective IR camera vendor.
How easy is it to integrate an IR camera into my system? Ethernet connectivity is the de facto standard for data communications in machine vision, and IR camera manufacturers design cameras that can plug-and-play as simply as possible. Meaning that at the very least, you should be looking at IR cameras that are GigE Vision and GenICam compliant.
What kind of software do I need to integrate and control the IR camera and analyze the images? If the IR cameras you purchase are plug-and-play friendly, then they should also be able to be supported by a broad range of third-party software. Unless you possess a great affinity for writing source code, you should be able to install the cameras and define complex pass/fail decisions, control digital I/O and communicate with programmable logic controllers (PLCs), PCs or human machine interfaces (HMIs) via a built-in deployment interface. Thermographic software is key to removing the burden of image transfers, data analysis and decision making from your already busy Ethernet network. Some cameras may come with theromographic software, but there’s also a good market for third-party software if you want more than what your vendor’s software has to offer.
What resolution do I need in my IR cameras? IR cameras come in a variety of resolutions. It’s important to choose the appropriate resolution for your application because too little can lead to disappointing results and too much can take a toll on your network (and is much more expensive). Generally, ultra-high detector resolutions are not needed in machine vision, so a typical focal plane array (FPA) would be 320 × 240 pixels.
Does your IR camera produce a temperature reading? You might not need it, especially if pass/fail can be determined simply by differentiating hot and cold. You can save on upfront costs if you don’t need a calibrated IR camera. However, if determining pass/fail is based on varying degrees (pardon the pun) of temperature deltas-which seems to be the case in a majority of machine vision applications, including food processing, packaging, electronics manufacturing and more-then make sure you get a calibrated IR camera.
How many IR cameras do I need? This is where you’ll want to rely on an expert to help you out. Depending on your site and your application, you may want to install a variety of cameras. Multi-sensor packages on high-performance pan/tilts offer IR and visible light cameras in a single installation where monitoring temperature and security are both priorities. Fix-mounted IR cameras allow you to survey a set area; the field of view being covered depends on the optics you select. There is no established configuration that automatically gives you complete visibility; every system is custom. You’ll want to work closely with your IR camera vendor to make sure you’re able to see everything you need, no matter the remoteness of some of your facilities.
Will an IR camera survive in the locations I need to install it? IR cameras are highly sensitive sensors. Manufacturers do the best they can to protect the cameras, but make sure you know the intricacies of your warranty before implementation. Also, make sure you know ahead of time if you need to invest in extra protection in the form of poles or environmental housing.
IR imaging is invaluable to machine vision. Once you see what an IR camera sees, you immediately recognize the benefits of this nonintrusive technology. Then after you see what an IR camera can do for your application, you’ll probably be a little angry with yourself for not implementing it sooner. That anger quickly dissolves after the first shutdown, equipment replacement or disaster is averted thanks to your IR camera. V&S
Tech TipsIR imaging involves pointing an infrared camera at a particular scene where you need to see more than what you can with the naked eye or through a visible light camera.
IR is emitted by all objects at temperatures above absolute zero, and the amount of radiation increases with temperature.
A properly calibrated IR camera can capture thermographic images of target objects and can provide accurate noncontact temperature measurements of those objects.