In the food industry, it's essential to carefully control the temperature of perishable goods throughout production, transportation, storage and sales. Repeated warnings about infections due to tainted and improperly cooked foods highlight the need for tighter process control. Since this almost always involves a human factor, food processors need tools that automate crucial operations in a way that helps minimize human error while holding down costs. Infrared (IR) thermography is such a tool.
Figure 1. IR camera of the type used for automated non-contact temperature measurements in food processing applications. Analog video outputs can be viewed on video monitors, and digital MPEG4 video outputs can be routed to a computer via Ethernet
The main elements of automated thermography are an IR camera (Figure 1) and associated software. They act as “smart” non-contact sensors to perform 100% inspections, measuring the temperature of cooked food items as they exit the cooking process. Measurement accuracy is typically ±2°C, they are easy to use, small, and can be positioned almost anywhere as needed.
IR cameras are also used to monitor cooking equipment, inspect package sealing, and improve efficiency in other food processing operations. Many IR cameras today have firmware and communication interfaces that enable their use in automated process control. Third-party software makes it easy to incorporate these tools into automated machine vision systems without the need for extensive custom written control code.
The use of IR cameras in food processing is growing for applications such as:
- Microwave cooked meats
- Oven baked goods
- Microwave drying of parboiled rice and other grains
- Inspecting ovens for proper temperature
- Proper filling of frozen entrée package compartments
- Checking integrity of cellophane seals over microwave entrées
- Inspecting box flap glue of overwrap cartons
- Monitoring refrigerator and freezer compartments
Obviously, some of these applications are more critical than others in terms of food safety. However, all these uses of thermography contribute to improved efficiency in the food processing industry.
Thermography for QA and Product Safety
IR thermography is first and foremost a quality assurance (QA) tool. Controlling the quality and safety of cooked meat products is an excellent use of this technology. A permanently mounted IR camera can record the temperature of, for example, chicken tenders as they exit a continuous conveyor oven (Figure 2). The objective is to make sure they are done enough (reach the right temperature), but not over-cooked and dried out.
Figure 2. Digital photo (left) and IR thermogram (right) of a chicken tender oven conveyor line.
Note the temperature scale in the thermogram of Figure 2. It shows that the chicken tenders just exiting the oven are at a temperature of about 150°F. (Camera firmware allows the creation of temperature scales with more intermediate divisions if desired.) Temperatures below 140°F pose safety risks from e-Coli and other harmful bacteria that may not have been killed. While higher temperatures provide a margin of safety, going above 170°F can dry out the product, degrading its texture, aroma, and taste. Reduced moisture content also represents yield loss on a weight basis.
Considerable research has been conducted to correlate surface temperatures obtained through thermographic measurements with product core temperatures taken with thermocouples or other contact sensors . For example, a Georgia Tech Research Institute (GTRI) Agricultural Technology Research Program study  found that the core temperature relative to the surface temperature of cooked chicken products depends on a number of factors. These include whether or not the chicken pieces have been breaded, how heavy that coating is, and whether the coating has been ruptured.
In addition, there is the possibility that some of the pieces may have been colder than others when they went into the oven, or may have gotten partially covered by another piece during handling or on the conveyor. In any case, pieces that come out of the oven with a temperature below or above the levels set in IR camera firmware will cause an alarm output to be triggered. This will also show up on a video or PC monitor so an operator can remove bad pieces from the conveyor.
Thermography for Microwave Precooking
GTRI also studied the use of microwave precooking to speed up subsequent cooking in a conventional conveyor oven. Their researchers found that by combining IR thermograms with visible-light images, they could avoid the non-uniform heating that sometimes occurs in microwave ovens. (Some IR cameras also have a built-in digital image camera that eliminates the need for a separate visible-light camera.) By examining both types of images, they combined information on product temperature, color, and size that led to the creation of a control algorithm that analyzes the data and allows adjustments to the microwave cooking system on the fly. This is more difficult to do in a conventional oven system.
Besides improving product quality and safety, overall throughput is increased with these thermography augmented cooking systems. An additional benefit is reduced energy costs.
In addition to cooked food inspections, IR thermography can also monitor the conveyor ovens. GTRI is in the process of using this to create a verification step and feedback loop to control oven temperature.
Another use of IR for conveyor ovens is monitoring temperature uniformity across the width of the cooking belt. If a heating element inside an electric oven fails, or you get uneven heating across an air impingement oven, one side of the product stream may be cooler. This can be quickly discovered with thermographic images and the IR camera’s alarm system. Quality inspections of this sort are much more difficult with conventional contact type temperature sensors. Thus, IR can help correct product variability and improve quality before a lot of product needs to be scrapped.
Even microwave oven heat patterns can be monitored with IR thermography. Figure 3 is a thermogram of the interior of a microwave oven showing horizontal mode heating.
Figure 3. Thermogram of the horizontal mode heat pattern in a microwave oven. A glass plate with a thin film of water was placed at a height of 8cm to detect the pattern.
In the final cooking process it’s still a good idea to periodically sample meat product core temperatures with thermocouples or other types of contact sensors. Even there, thermographic images and temperature measurements can help inspectors give more attention to products likely to be outside the acceptable temperature range.
Other Baked and Roasted Products
Recent salmonella problems seem to suggest that manufacturers of roasted nut products might benefit from thermographic inspections. As alluded to above, nuts roasted in heated air impingement ovens may be prone to uneven cooking. An IR camera can be used to inspect the roasted nuts and the oven for proper temperature.
Another possible use of IR cameras is combining temperature measurements with pattern matching for baked goods such as cookies. The temperature measurements check for doneness, while the camera’s pattern matching capabilities can make sure that finished pieces conform to the required shape of the product.
Software is available that allows a thermographic vision system to learn and locate objects and patterns in the images One application for pattern matching is in the production of frozen entrées. IR machine vision can use pattern recognition software to check for proper filling of food tray compartments.
A related application is automated 100% inspection of the heat-sealed cellophane cover over finished microwave entrées. IR cameras can see heat radiating from the lip of the container where the cellophane heat-seal is formed. The temperature along the entire perimeter of the package can be checked by using the camera’s thermographic image with machine vision software. This type of program matches the geometric pattern in the image and its temperatures against the temperatures in a pattern stored in computer memory.
This is accomplished by programming the specific pixels in an image to be used by the software for this purpose. Typically, the software looks for minimum, maximum, or delta values that tells the equipment operator if the package passes inspection. An added function in such a system could be laser marking of a poorly sealed package so it can be removed at the inspection station.
An issue affecting product safety indirectly is the integrity of cartons that overwrap and protect food containers. One of the most cost effective ways of sealing overwrap cartons is to use heated glue spots on the carton flaps. In the past, the integrity of the spot gluing was determined by periodically doing destructive testing on several samples. This was time-consuming and costly.
Since the glue is heated, an IR camera can “see” through the cardboard to check the pattern and size of the applied glue spots (Figure 4.) The camera can be set up to look at predefined areas of the flaps where glue should be applied, and verify spot sizes and their temperatures. The digital data collected is used for a pass/fail decision on each box, so bad boxes can be immediately removed from the production line. The data is automatically logged into the QA system for trend analysis, so a warning can be generated if an excessive number of boxes begin to fail.
Figure 4. Thermogram of box flap glue (left) and screen capture (right) of the software algorithm for making pass/fail decisions using the National Instruments Vision Builder for Automated Inspection (VBAI) package.
Yet another application for thermography can be found in bottle filling operations. Although this is seldom a product safety issue, it does affect yield and compliance with regulations. As shown in Figure 5, IR imaging can be used to monitor bottle filling operations. Different areas on the bottle can be defined and used to trigger an alarm, and then remove bottles that are over- or under-filled. Thermography is a better alternative to visible light cameras when a bottle or jar is made of dark colored glass or plastic.
Figure 5. Thermogram of a bottle filling application, showing a low level in the second bottle from the left.
Automating Thermographic Measurements
Application software currently available provides functions that support a wide variety of food processing applications, and complement the functions built into IR cameras. The imaging tools and libraries in these packages are hardware- and language-independent, making it easy for food processing engineers to quickly implement thermographic monitoring and control systems.
IR cameras themselves provide the user with different operating modes that support correct temperature measurements under various conditions. Two functions commonly found in these cameras are a spotmeter and area measurements (Figure 6). The spotmeter finds the temperature at a particular point. The area function isolates a selected area of an object or scene and usually provides the maximum, minimum, and average temperatures inside that area. The temperature measurement range typically is selectable by the user. As an adjunct to the temperature range selection, most cameras allow a user to set up a color scale or gray scale to optimize the camera image, as shown on the right side of Figure 6.
Figure 6. Area (Ar) and spot (Sp) non-contact temperature measurements of hamburgers on an oven conveyor line.
In conveyor oven applications, the area function is typically used, because pieces of cooked product are often randomly located on the conveyor. The camera can be programmed to find and measure the minimum and maximum temperatures within the defined area. If one of those setpoint temperatures were to fall outside the user-defined limits, an application program running on a PC or PLC would instantly trigger an alarm, alerting the operator to check the thermographic image on a video monitor or PC to find and remove the bad product, and/or adjust the cooking temperature.
All this is made possible with analog and digital interfaces on the IR camera. For cameras with an Ethernet connection, such as the FLIR Model A320/325 Series (Figure 1), digitally compressed (MPEG-4) streaming video is available for monitoring on a PC screen. In addition to local monitoring, images and alarms can be sent to a remote PC via ftp and SMTP (email) protocols. Temperature data such as minimum, maximum, and average for specific pieces of food or the camera’s field of view (FOV) as a whole can be collected for trend analysis and statistical process control (SPC) purposes.
Figure 7. Simplified system diagram for a food conveyor application.
A simplified block diagram of conveyor monitoring is shown in Figure 7. This could be representative of a conveyor oven application. One IR camera may be adequate for many applications, or an IR camera may be combined with a visible light camera to record other target object attributes.
In the case of local monitoring, an IR camera’s digital I/O can be used to directly trigger an alarm device without additional software. However, food processing often benefits from higher level analytics that are available in third party software that runs on a PC. These out-of-the-box solutions do not require the writing of application source code. By adhering to commonly used machine vision interface standards such as GigE Vision™ and GenICam™, a wide range of functionality is supported by this software.
Although thermographic cameras and associated software can recognize the temperature, size, shape, and relative location of target objects, camera positioning and protective enclosures may be an issue in conveyor line applications. Placing a camera directly over a conveyor line generally is not acceptable unless a protective shield is placed between it and the food to protect against possible contamination from the camera. Conventional glass and plastics that are transparent to visible light greatly attenuate IR radiation, making accurate temperature measurements nearly impossible.
One way to avoid this problem is to position the camera at an angle to the conveyor and a few feet away from it. This requires a calibration adjustment, but is a viable option. Another option is to place the camera in a food compliant enclosure with a special window that is transparent to IR energy, thereby allowing the camera to be located directly over the conveyor. These enclosures can be air or water cooled to avoid heat damage to the camera.
In automated machine vision systems, IR cameras are frequently combined with visible light cameras to record other target object variables. In the case of food processing, particularly with baked goods, color is another indication of proper cooking and product quality. With the type of software mentioned earlier, this combination of cameras can also monitor a wide range of size, shape, and temperature variables for highly sophisticated monitoring and control, plus 100% inspection.
Jason Styron is an automation business development manager with FLIR Systems, Inc. He earned his Electrical Engineering degree from Auburn University, and has 12 years of experience in industrial automation. His last five years have been focused on infrared thermography and its application in automated machine vision applications. He can be reached at Jason.Styron@FLIR.com.
1. Ibarra, Juan G.; Tao, Yang; Xin, Hongwei (in collaboration with Tyson Foods), “Combined IR imaging-neural network method for the estimation of internal temperature in cooked chicken meat”, Optical Engineering, Vol. 39, No. 11, November 2000.
2. Sanders, Jane M., “Seeing the Unseen: Infrared computer vision system could help make meat products safer, tastier, and less costly to produce”, PoultryTech: Volume 19, Number 1, Spring 2007.
3. FLIR Systems, Inc., “IR Automation Guidebook”, 2008, Chapter 4, Combining Machine Vision and Temperature Measurements, pp 27-28.