Thermal Imaging

With Thermal Cameras, Image Matters

Adoption of thermal cameras has been growing rapidly, and new vendors are coming onto the field at a rapid clip. This can make getting the camera or collection of cameras you need more problematic rather than less so.

The simple fact is that not all thermal security cameras are created equal. Simply comparing manufacturers’ specification sheets won’t give you the information you need to get the most effective imagers for your money.

For many security professionals, the process of evaluating thermal cameras for purchase or recommendation is a new endeavor that exposes them to a whole new set of specifications and performance parameters that they are unfamiliar with. This unfamiliarity makes them vulnerable to the vagaries of slick marketing literature.

So, how can you get the right cameras? Start by asking the right questions.

How is a thermal camera’s resolution measured, and why do I care?

The detector is the heart of any thermal security camera. It’s the part that gathers the infrared energy and allows the creation of an image made from this energy. A thermal camera’s detector plays the same role as the CCD detector chip in a standard video camera—whether it’s on a pole outside of a nuclear facility or in your handi-cam at home.

The detector’s resolution is the number of individual detector elements found on that chip, usually measured in horizontal and vertical dimensions. The low-end options for thermal resolution typically offered are the 160x120 detector or the 320x240 detector formats, but the standard is quickly becoming the 640x480 detector. There’s good reason for this.

A detector’s resolution is the first vital element in determining a camera’s ability to generate a high-quality image. The more detector elements a detector has— meaning, the higher its resolution—the more energy will be gathered, and the more detail you’ll be able to see in the image.

Also, if you were to compare a camera with 640x480 resolution and a camera with 320x240 resolution that use the same size of lens, you’ll find that the 640’s angular field of view will actually be wider, yet will also detect threats from farther away. In the real world, this means that you’ll be able to cover the same amount of area with fewer cameras using 640 resolution and still be able to detect intruders from farther away. In other words: lower cost with better performance.

Let’s take two national thermal cameras that use the same uncooled VOx detectors and 35 mm lenses as an example. The 320 camera will have a 13 degree x10 degree FOV, while the 640 camera will have an 18 degree x14 degree FOV. This will give you a nearly 40 percent increase in coverage area, while still being able to detect a person from nearly 50 percent farther away. The tactical and economic benefits of increased resolution are measurable and undeniable.

We’re so used to hearing about five-, 10- and 12-megapixel resolution in standard digital cameras that 640 sounds kind of puny. But it’s important to note that a 640 detector is the largest detector that is economically viable for commercial thermal security cameras. High-definition megapixel thermal cameras are available, but these are used for long-range military systems or scientific research and typically quite expensive.

But resolution is only the first step to creating a good image with lots of detail and solid contrast that will get the most out of an analytics package. Another important factor is what is often called pixel size or pixel pitch.

The detector’s pixel pitch is a specification that should be readily available from the manufacturer, as it is an important piece to the puzzle in predicting image quality and range performance. Typically measured in micrometers, or “microns,” when you’re looking at a camera’s pixel pitch, keep in mind that lower numbers are better— the smaller the pixel pitch, the more image detail you’ll get in a smaller package.

This time, let’s compare uncooled VOx cameras with 100 mm f/1.6 lenses; one camera has 25 micron pixels, and the other has 17 micron pixels. All other things being equal, the 17 micron camera will detect a person from more than 22 percent farther away than the camera with 25 micron pixels.

Just as important, however, the 17 micron camera will produce more detailed, higher contrast images that will get better results from analytics and VMD packages.

These technical factors aren’t enough to maximize a thermal camera’s performance. Technical factors such as detector type, lens design, detector resolution and pixel pitch will make sure you get the most thermal energy into the system’s electronics as possible, but what the camera does with that information is vitally important as well. That’s where image processing comes in.

A thermal camera’s internal image processing software can help bring out object edges and enhanced details that can make the difference in final image quality.

Image quality—isn’t it really just pretty pictures? No.

As with the technical attributes mentioned earlier—detector type, f/number, resolution, pixel pitch, and image processing—the benefits of what may be superficially regarded as a subjectively better image have objective and measurable benefits.

First let’s look at a couple of examples of better image quality. The pair of images on the top left of this page were taken with two leading thermal security cameras at 6 p.m. Both have 320 resolution, and they use similar lenses—a slight difference in the angular FOV can be noticed—and both are using their “out of the box” image settings. No optimization was done to either camera, and neither of the images have been manipulated in any way other than to re-size them for publication.

The image on the left has greater contrast, shows more small details, and is in focus throughout the entire depth of the image.

These points are notable because this is one of the times of day that can be most challenging for thermal imagers. Remember that thermal cameras don’t just make pictures from heat; they make pictures from tiny differences in heat.

There are two times of the day in which items within an image are most likely to have the smallest differences in temperature (isothermal): just before the sun goes down, when things have been soaking up the sun’s rays all day and have reached a critical point of solar loading, and in the middle of the night, when everything has radiated off its stored energy and cooled to a similar degree. Both conditions are called points of “thermal crossover.”

The camera that can keep imaging well during periods of thermal crossover will continue to give you solid coverage in the most challenging environmental conditions.

The other side of the coin is more dramatic. The second pair of images at the top of page 50 (the ones on the right) were taken at midnight, when things have become more uniformly cool.

The camera that produced the lefthand images in both pairs will outperform the other camera consistently in a number of important areas:

Threat detection. Detection by any means—direct human observation or alarms through VMD—will be dramatically improved with the better image quality shown by the images on the left of each pair.

Analytics performance. Better contrast and scene detail gives you better edge differentiation, better segmentation and better classification.

Range performance. When viewed with the camera on the left, small objects and small scene features have higher contrast and sharper edges, making them more readily recognizable across all distances from the camera.

Inclement weather. Fog, haze and precipitation of all forms reduce thermal contrast. Cameras with superior image contrast in these lowcontrast conditions will continue to provide optimal performance.


Which Is Better, Cooled or Uncooled?

Thermal security cameras are either cooled or uncooled, referring to whether the infrared detector at the heart of the camera’s sensor needs to be cooled to cryogenic temperatures in order to create an image. Which type of imager is better depends largely on the needs of the specific application; each has advantages and weaknesses.

Cooled cameras are more sensitive to small differences in scene temperature than are uncooled cameras, meaning they can see smaller objects from farther away, making cooled cameras more suitable for extremely long-range imaging in low-contrast environments. Because of their increased sensitivity, cooled cameras are best suited for applications such as critical infrastructure security and border security in which the detection of human targets beyond two miles is critical.

Uncooled thermal cameras—as the name implies—do not use cryogenic cooling. The most popular uncooled thermal security cameras use uncooled detectors called Vanadium Oxide (VOx) microbolometers. Uncooled sensors are typically sensitive to LWIR energy. Uncooled detectors are manufactured in fewer steps than those used in cooled sensors, use less expensive vacuum packaging, and, most significantly, don’t require costly cryocoolers.

Almost all commercial security applications opt for uncooled thermal cameras because they cost less, have similar short- to mid-range performance, and have much longer service lives than cooled cameras under similar operating conditions, enabling continuous operation.

How Does the Camera’s Lens Impact System Cost?

It has to do with another crucial lens parameter, the f/number. The f/number determines the light-gathering power of the lens and therefore affects the sensitivity of the camera system.

The f/number of an optical system is the ratio of the focal length of the lens to the diameter of the front lens element. An f/2 lens with a 500 mm focal length must therefore have a 250 mm diameter front lens element.

As the focal length of a lens increases, the diameter of the front lens element must be increased to keep the system f/ number constant. An uncooled camera must run at a low f/number—typically 1.4-2—to have sensitivity comparable to that of a cooled camera. Higher f/numbers reduce uncooled camera sensitivity.

These factors lead to the conclusion that short- to extreme medium-range imaging can usually be done more cost-effectively with uncooled thermal security cameras, while their cooled counterparts are the best solutions when long-range imaging performance is called for.

There’s no denying that the acquisition cost of a thermal camera is greater than that of a CCTV camera. But, by paying attention to the details of any thermal security cameras you are considering, you can actually get a technology that can do more for less.


This article originally appeared in the July 2011 issue of Security Today.


Featured Cybersecurity


New Products

  • Camden Door Controls ‘SER” Surface Boxes and Extension Rings

    Camden Door Controls ‘SER” Surface Boxes and Extension Rings

    Camden Door Controls has introduced new ‘SER” surface boxes and extension rings that provide a complete solution for new construction. In addition, they provide a simple and robust solution when replacing round wired and manual push plate switches with either Camden’s wired or wireless SureWave™ no-touch switches or Kinetic™ no-battery wireless switches. 3

  • Videoloft Cloud Video Surveillance VSaaS Solution

    Videoloft Cloud Video Surveillance VSaaS Solution

    Videoloft focuses on transforming traditional professional surveillance systems into cloud connected solutions via the Videoloft Cloud Adapter. 3

  • HID Signo Readers

    HID Signo Readers

    HID Global has announced its HID® Signo™ Biometric Reader 25B that is designed to capture and read fingerprints in real-world applications and conditions. 3