Thermal Imaging Specs Explained: NETD, Resolution, and Lens Size Demystified
Chris spent three weeks reading thermal imaging specs before buying anything. By week two, he'd filled a notebook with numbers and was more confused than when he started.
384×288. 12μm. NETD ≤40mK. f/1.0. 50Hz. 1.43-inch AMOLED. Pixel pitch. Detection range. Recognition range. Every spec sheet introduced three new terms that required their own research, each of which led to a forum debate where experienced users argued past each other using jargon that assumed knowledge he didn't have.
"I eventually just bought the most expensive one I could justify and hoped it was right," he said. "There has to be a better way."
There is. And it doesn't require an engineering degree.
Thermal imaging specifications are genuinely meaningful — the numbers tell you real, predictable things about how a device will perform in the field. But the way they're presented assumes you already understand the physics, and the marketing around them is designed to highlight the impressive-sounding numbers regardless of whether those numbers matter for your use case.
This guide translates every spec that actually matters into plain English, explains what it means in field conditions, and tells you what a good spec looks like versus a mediocre one at each price tier. By the end, you'll be able to read a thermal imaging spec sheet and know, immediately, what you're actually getting.
Why Thermal Imaging Specs Are So Confusing
Part of the confusion is structural. Thermal imaging specifications come from military and industrial applications where the audience already has the conceptual framework to interpret them. When consumer devices entered the market, manufacturers carried over the professional specification format without providing the translation layer.
The other part of the confusion is intentional. Manufacturers have every incentive to highlight the specs their device does well and minimize or omit the specs where it underperforms. A device with excellent NETD sensitivity but poor display resolution will lead with the NETD number. A device with a large, vivid display but limited sensor resolution will showcase the display specs.
Understanding the full picture requires knowing which specifications are actually diagnostic for performance and which are secondary or frequently misrepresented.
There are five specifications that matter most in a thermal imaging device for outdoor observation and field use. Everything else is either secondary or a derivative of these five.
In order of importance: NETD sensitivity, sensor resolution and pixel pitch, lens specifications, display quality, and refresh rate.
Specification #1: NETD — The One That Matters Most
NETD stands for Noise Equivalent Temperature Difference. It's the specification that generates the most confusion and is the most important single number in a thermal imaging spec sheet.
What it actually measures:
NETD tells you the smallest temperature difference the sensor can reliably detect and display as distinguishable from background image noise. It's measured in millikelvin (mK) — thousandths of a degree. A sensor with NETD of 40mK can distinguish thermal differences of 0.04°C. A sensor with NETD of 80mK requires a temperature difference of 0.08°C before it can separate signal from noise.
Lower numbers are better. This is the most important thing to remember about NETD: lower is more sensitive.
Why it matters in the field:
The core function of a thermal device is detecting the temperature difference between a warm-bodied subject (an animal, a person) and its surrounding environment. The greater that temperature difference, the easier the detection. The smaller the difference, the harder — and the more NETD sensitivity matters.
Consider two scenarios:
Winter night at 30°F: A deer at 101°F is 71°F warmer than the ambient environment. Even a device with poor sensitivity (80mK or higher) produces a vivid, clearly visible deer against the cool background. The temperature difference is large enough that almost any thermal sensor can resolve it.
Summer night at 78°F: The same deer is now only 23°F warmer than the ambient environment. The thermal contrast is dramatically lower. A device with 80mK sensitivity may produce a soft, ambiguous image where the deer barely stands out from the background. A device with 40mK sensitivity resolves the same scene with meaningful contrast — the deer is visible and readable even against the warm summer background.
This is the practical consequence of NETD: it determines whether your thermal device works well in warm ambient conditions, not just in cold ones.
NETD values and what they mean:
≤25mK — Military and professional grade. Found in high-end devices above $2,000. Exceptional sensitivity in all conditions.
≤40mK — The threshold for serious field performance in variable weather. Devices in the $500–$1,200 range that deliver this spec provide reliable warm-weather performance. This is the minimum spec worth targeting for any serious outdoor use.
50mK — Acceptable in cold-weather conditions. Warm-weather performance is noticeably reduced. Common in devices priced $300–$500.
60mK and above — Adequate in cold ambient conditions only. Summer use with small or partially concealed subjects will produce unreliable results. Common in budget devices under $300.
What NETD is not:
NETD is not a direct measure of image quality or visual sharpness — a device with excellent NETD but poor sensor resolution can still produce a blurry image. And NETD is a sensor-level specification measured under controlled laboratory conditions; real-world performance can vary based on the optical system, processing algorithms, and sensor manufacturing quality. A device claiming ≤40mK should be evaluated alongside other specifications, not taken in isolation.
Specification #2: Sensor Resolution and Pixel Pitch
Sensor resolution is the number of individual thermal detector elements in the sensor array, expressed as a grid: 256×192, 384×288, 640×480. Pixel pitch is the physical size of each detector element, expressed in micrometers (μm).
These two specifications work together to determine image spatial resolution — essentially, how much fine detail the sensor can capture.
Understanding sensor resolution:
256×192 = 49,152 total pixels 384×288 = 110,592 total pixels 640×480 = 307,200 total pixels
Higher numbers mean more total detector elements capturing the scene. A 384×288 sensor captures 2.25 times more spatial information about the scene than a 256×192 sensor, and that information translates to finer detail in the image.
The practical consequence: at 200 meters with a 256×192 sensor, a deer is represented by fewer pixels than at the same distance with a 384×288 sensor. Fewer pixels means less shape definition — the difference between "I can tell that's a deer by its body shape" and "I can tell that's a deer because there's a warm blob roughly that size."
Understanding pixel pitch:
Pixel pitch is the physical size of each detector element. Common values are 12μm, 17μm, and in older or budget sensors, 25μm or larger.
Smaller pixel pitch (12μm) means smaller pixels, which means more pixels can fit in the same sensor area. This produces higher spatial resolution for the same sensor size.
Larger pixel pitch (17μm) means larger pixels. Each individual pixel collects more thermal radiation than a smaller pixel of the same sensitivity — there's an argument that larger pixels can be more sensitive to very faint signals. However, the spatial resolution penalty of larger pixels is significant at the distances where observation detail matters.
The 12μm vs 17μm real-world difference:
A 384×288 sensor at 17μm pixel pitch and a 384×288 sensor at 12μm pixel pitch have the same number of pixels, but the 12μm sensor provides finer angular resolution — each pixel covers a smaller angle of the field of view, meaning the image resolves more fine detail at equivalent distance.
The 12μm sensor at 384×288 resolves approximately 42% more spatial detail per unit field of view than a 17μm sensor of the same resolution. At 200 meters, this is the difference between clearly resolving a coyote's ear shape and body proportion versus seeing a roughly coyote-sized warm shape.
What about AI super-resolution?
AI super-resolution processing takes the output of a lower-resolution sensor (typically 256×192) and applies trained algorithms to reconstruct a higher-resolution image — typically displaying equivalent quality to a native 384×288 sensor at close to medium distances.
The technology works by analyzing sub-pixel thermal gradients captured by the sensor and using AI inference to interpolate additional image detail that the native resolution couldn't display directly. The result is displayed at 384×288 equivalent quality from a 256×192 sensor, at a hardware cost that allows the device to be priced significantly lower than a native 384×288 device would be.
At typical observation distances (under 250 meters), well-implemented AI super-resolution produces images that are meaningfully difficult to distinguish from native 384×288. At extreme distances (300+ meters), the native sensor's advantage over AI upscaling becomes more apparent. For most civilian outdoor use cases, AI super-resolution in a quality implementation represents genuine value rather than a gimmick.
Resolution guide in plain terms:
256×192 native, no AI processing — Close-range use only. Adequate for 50–100m observation where subjects fill a significant portion of the frame.
256×192 with quality AI super-resolution — Effective to 200–250m for typical wildlife and field observation. Good value at mid-range price points.
384×288 at 12μm — Full-capability device for observation to 300m with recognizable detail. The spec target for serious field use.
640×480 — Long-range observation beyond 400m. Professional and premium tier. Necessary only for specialized applications.
Specification #3: Lens — Focal Length and Aperture
The lens is the optical element that collects thermal radiation from the scene and focuses it onto the sensor. Two numbers describe a thermal lens: focal length (in millimeters) and aperture (expressed as f/number).
Focal length determines field of view and detection range:
Longer focal length = narrower field of view, higher apparent magnification, better long-range detection. Shorter focal length = wider field of view, lower apparent magnification, better scanning capability.
A 35mm lens produces a narrow field of view — typically 7–8 degrees horizontal — that is well-suited to observing specific subjects at distance. The apparent magnification brings subjects at 300 meters into clear view.
A 15mm lens produces a wider field of view — typically 10–12 degrees horizontal — that covers more of the landscape per sweep. Subjects at 300 meters appear smaller than they would through a 35mm lens, but you're covering 50–60% more landscape in each scan. This is better for detection (finding animals across a wide area) and less efficient for identification at extended range.
The field-use implication: a device with a 15mm lens is better for systematic scanning to locate subjects. A device with a 35mm lens is better for observing specific subjects you've already located. For observers whose primary challenge is detection across a wide landscape, the 15mm lens is more useful. For observers whose primary challenge is detail observation of subjects at distance, the 35mm lens is more useful.
Aperture (f/number) determines thermal radiation collection:
The f/number describes the relationship between the lens focal length and the aperture diameter. Lower f/numbers indicate a larger effective aperture relative to focal length — more thermal radiation collected per unit time.
f/1.0 — Standard for quality mid-range thermal devices. Good light collection. f/0.9 — Approximately 23% more radiation collection than f/1.0. The spec found in devices that prioritize warm-weather and low-differential performance. f/1.2 or higher — Reduced radiation collection. Found in budget devices or specialized designs. Noticeable performance difference in warm ambient conditions.
The practical consequence of aperture: on a warm night when the temperature differential between an animal and its environment is 20°F rather than 60°F, an f/0.9 lens maintains usable image contrast where an f/1.2 lens produces a flatter, lower-contrast image. The aperture directly affects warm-weather performance in the same direction as NETD — both specifications matter more when conditions are challenging than when conditions are ideal.
The f-number and NETD interaction:
These two specifications work together. A device with excellent NETD sensitivity (≤40mK) but a slow aperture (f/1.2) will underperform its NETD spec in warm conditions because the lens is limiting the radiation reaching the sensor. A device with average NETD (50mK) but an excellent aperture (f/0.9) may perform better in warm conditions because the faster lens delivers more signal to the sensor.
Evaluating NETD and aperture together provides a more complete picture of warm-weather performance than either number alone.
Specification #4: Display — Screen Size, Type, and Resolution
The display is the final step in the image chain, and it's the step where many devices lose the detail their sensor and lens worked to capture.
Display types:
OLED (Organic Light-Emitting Diode) — Each pixel emits its own light, producing deep blacks (pixels turn off completely rather than blocking a backlight) and high contrast. The black-to-white contrast ratio is relevant to thermal imaging because thermal images fundamentally represent contrast between warm and cool objects. OLED and AMOLED displays show this contrast more vividly than LCD alternatives.
AMOLED (Active Matrix OLED) — A specific OLED architecture that allows individual pixel control at high speed. The result is faster response time and the ability to produce very large, high-quality screens efficiently. AMOLED is the display type found in most quality smartphones and increasingly in premium thermal devices.
LCD — Cheaper to manufacture, adequate quality for many applications, but inherently limited in black-level performance because a backlight is always on and blocking it is never complete. In thermal imaging, where black-level matters for the display of cold backgrounds, LCD produces a slightly washed-out image compared to OLED at equivalent resolution.
Display size and format:
Traditional thermal monoculars use a small eyepiece display (typically 0.23 to 0.39 inches diagonal) magnified through an eyepiece lens. The eyepiece format requires pressing the device against your eye to observe — the same viewing geometry as a conventional optical device.
Some devices use a larger direct-view display — a screen large enough to observe at arm's length, like a smartphone, rather than pressed to the eye. The GTGUARD Hawkeye AI15's 1.43-inch AMOLED screen is the largest display in its price category, providing a significantly different viewing experience from traditional eyepiece designs.
The practical difference: eyepiece viewing keeps the observer's face pressed to the device for extended sessions — more stable, more focused, but physically demanding for sustained periods. Direct-view display allows comfortable arm's-length observation without facial contact — better for extended sessions, casual scanning, and situations where keeping the device at eye level isn't convenient or comfortable.
Display resolution:
Display resolution should be matched to sensor resolution. A high-quality sensor outputting to a low-resolution display loses the detail the sensor captured. The display resolution doesn't add detail the sensor didn't capture — but low display resolution can subtract detail the sensor did capture.
Quality mid-range devices use displays of 1024×768 (for eyepiece format) or equivalent quality in direct-view formats. The AMOLED display in the AI15 at 466×466 pixels provides sharp rendering within its format that preserves the AI-enhanced image quality without bottlenecking at the display stage.
Specification #5: Refresh Rate
Refresh rate is how many times per second the thermal image is updated, measured in Hertz (Hz). Common values are 25Hz (also called 25fps) and 50Hz (50fps).
What the difference looks like:
At 25Hz, the image updates 25 times per second. For stationary or slow-moving subjects, this is perfectly adequate — there's no visible difference between 25Hz and 50Hz when observing a deer browsing slowly 100 meters away.
For fast-moving subjects — a coyote sprinting across a field, a bat working the airspace above a pond, a bird flushing from cover — 25Hz produces a slightly stuttered, less fluid image that can make tracking and identification harder. The motion blur and frame gap of 25Hz is perceptible during rapid movement.
50Hz produces a smooth, cinematic image of moving subjects at the equivalent of 50 frames per second. Fast-moving animals track fluidly, and the continuous image stream makes maintaining visual contact with moving subjects significantly easier.
When refresh rate matters most:
For wildlife observation of mammals at typical viewing distances, 25Hz is adequate for most scenarios. 50Hz is the meaningful upgrade for observing fast-moving subjects — birds, bats, running coyotes — and for situations where tracking movement across the frame is important.
For applications involving moving subjects as a primary use case, 50Hz is worth prioritizing. For primarily stationary or slow-moving observation (deer at a field edge, wildlife at a water source), 25Hz is functional.
Putting It Together: Reading a Real Spec Sheet
With these five specifications understood, reading a thermal device spec sheet becomes a structured analysis rather than a confusing list of numbers. Here's how to evaluate any thermal imaging device in five questions:
Question 1: What is the NETD? ≤40mK or better: suitable for warm-weather and year-round use. 50mK: cold-weather adequate. 60mK+: cold-weather only for serious use.
Question 2: What is the sensor resolution and pixel pitch? 384×288 at 12μm: full-capability. 256×192 with quality AI processing: mid-range capable. 256×192 native only: close-range use.
Question 3: What is the lens focal length and aperture? 15mm for wide-field scanning. 35mm for extended-range observation. f/0.9–f/1.0 for warm-weather performance; f/1.2+ acceptable in cold-weather-only use.
Question 4: What is the display type and size? OLED or AMOLED: high contrast. LCD: adequate. Display resolution should match or exceed sensor resolution.
Question 5: What is the refresh rate? 50Hz: preferred for moving subjects. 25Hz: adequate for stationary or slow-moving observation.
Recommended Device: GTGUARD Hawkeye AI15 Thermal Monocular — $529
Having established the framework for evaluating thermal imaging specifications, applying it to a specific device produces a concrete assessment rather than marketing-driven enthusiasm.
The GTGUARD Hawkeye AI15 at $529 is the device that comes up consistently when the spec-to-price ratio is the primary evaluation criterion. Here's the spec assessment using the framework above.
NETD: ≤40mK
The AI15 hits the target specification for year-round, warm-weather capable thermal performance. At ≤40mK, it resolves the warm-ambient temperature differentials of summer wildlife observation — the scenario where budget thermal devices with 60mK+ sensitivity produce flat, unusable images. This is the most important specification on the AI15's sheet, and it's right.
<cite index="18-1">Crystal Clear: NETD≤40mk sensitivity detects the slightest heat variances.</cite> At $529, delivering ≤40mK sensitivity is the specification that places the AI15 in a different performance category from devices priced similarly that don't achieve this number.
Resolution: 256×192 with AI super-resolution to 384×288 display quality
The AI15 uses a 256×192 base sensor enhanced by AI super-resolution processing — the same technology discussed in Specification #2. <cite index="18-1">The Hawkeye AI15 utilizes cutting-edge AI Resolution Technology to upscale a standard 256×192 sensor into a stunning 384×288 visual experience.</cite>
The implementation quality determines whether this is genuine value or marketing language. At distances under 200 meters — which describes the majority of wildlife observation at campgrounds, backyards, and forest edges — the AI-enhanced image is meaningfully sharp and detail-rich. At 200–300 meters, it produces usable images for detection and rough identification. Beyond 300 meters, a native 384×288 sensor has a real advantage.
For the use cases the AI15 is designed for — wildlife observation, outdoor recreation, property monitoring at typical civilian distances — the AI processing delivers on its specification.
Lens: 15mm at f/0.9
<cite index="18-1">Pro-Grade Optics: 15mm lens paired with AI-enhanced processing delivers the detail of a high-end device at a fraction of the cost.</cite> The 15mm focal length produces a wide 8.8°×7° field of view — well-suited to the scanning applications where detection across a landscape is the primary goal. The f/0.9 aperture is, as discussed, the fastest in GTGUARD's lineup and provides approximately 23% more thermal radiation collection than f/1.0 alternatives.
For wildlife observers whose primary challenge is finding animals across a broad landscape rather than observing specific subjects at extreme range, the 15mm wide-field lens is the right specification. For observers who primarily work at known distances with specific identified subjects, a longer focal length would provide more apparent magnification.
Display: 1.43-inch AMOLED
<cite index="18-1">Enjoy a dramatically larger viewing area than standard 0.23" or 0.27" screens — up to 20× bigger — making it easier to spot heat signatures and track movement with clarity and comfort, even during extended use.</cite>
The 1.43-inch AMOLED direct-view display is the AI15's most distinctive physical feature. The AMOLED technology provides the deep blacks and high contrast that thermal imaging benefits from. The size — 1.43 inches versus the 0.23–0.39 inches of conventional eyepiece displays — changes the observation experience fundamentally: comfortable at arm's length for extended sessions, shareable with another observer, and practical for scanning without pressing the device to your face for hours.
This display format suits extended observation sessions — watching a meadow for 90 minutes, scanning a campsite periodically throughout the night — where the physical comfort of arm's-length viewing matters.
Refresh rate: 50Hz
<cite index="18-1">Fluid Motion: 50Hz high frame rate ensures zero lag while tracking fast-moving targets.</cite> The AI15's 50Hz refresh rate handles fast-moving wildlife — bats, birds, running coyotes — with smooth, fluid imaging that lower-rate devices cannot match.
Additional specifications worth noting:
Detection range of 1,500 meters — well beyond the practical recognition range but confirming the device's optical system quality. Weight at 335 grams — light enough for extended carry. Battery life at 10+ hours from the built-in cell — a full observation session without battery management.
The honest AI15 assessment:
The AI15 is the right device for an observer whose primary use is wildlife detection and observation at distances under 200 meters, who values the wide-field scanning that a 15mm lens provides, and who will benefit from the physical comfort of the large AMOLED direct-view display over traditional eyepiece designs.
It's not the right device for an observer whose primary use is extended-range observation beyond 250 meters, who needs the narrow field of depth that a 35mm lens provides for specific subject detail at long distance, or who specifically prefers the traditional eyepiece-to-eye observation geometry.
At $529, the AI15's NETD ≤40mK, 50Hz refresh, f/0.9 aperture, and AMOLED display represent the specifications that matter most, at values that justify the price against alternatives.
Common Spec Mistakes to Avoid
Mistake 1: Prioritizing detection range over recognition range.
Detection range (how far the device can confirm something is there) and recognition range (how far you can tell what it is) are fundamentally different numbers. A device claiming 1,500m detection range may have a recognition range of 300m — useful for wildly different purposes. Always ask what the recognition range is for the target size you'll typically be observing.
Mistake 2: Assuming higher magnification is always better.
Digital zoom increases displayed image size at the cost of image quality. Optical magnification from a longer focal length lens provides genuine detail. A device with 8× digital zoom from a 15mm base is not the same as a device with 4× from a 35mm base. The focal length and sensor resolution determine actual detail capability; digital zoom numbers are marketing.
Mistake 3: Ignoring NETD and focusing on resolution.
Sensor resolution without adequate NETD sensitivity is incomplete. A 384×288 sensor with 80mK NETD will produce worse images in warm conditions than a 256×192 AI-enhanced sensor with ≤40mK NETD. Both numbers matter; NETD is frequently omitted or minimized by manufacturers whose devices don't perform well on that metric.
Mistake 4: Not asking what the lens aperture is.
Lens aperture is frequently missing from consumer thermal spec sheets — it's not a number manufacturers who use slow lenses want highlighted. Always look for the f/number. f/1.0 or better is the target; f/1.2+ at the same price tier represents a meaningful performance compromise.
Mistake 5: Choosing a device for specs you don't need.
640×480 resolution is genuinely superior to 384×288. It's also $1,500+ more expensive in the devices that implement it well. If your observation scenarios are consistently under 200 meters, the performance advantage of 640×480 over 384×288 AI-enhanced is marginal and doesn't justify the price premium. Match specifications to use case; don't buy specifications you won't use.
The Spec Sheet Checklist
Before purchasing any thermal imaging device, verify these five items from the spec sheet:
- NETD: ≤40mK for warm-weather capable. 50mK for cold-weather adequate.
- Sensor resolution + pixel pitch: 384×288 at 12μm native, or 256×192 with confirmed quality AI processing. Look for the pixel pitch number — its absence is a warning sign.
- Lens focal length and aperture: Match focal length to your use case (15mm wide for scanning, 35mm for extended-range detail). Look for f/1.0 or better.
- Display type and size: OLED or AMOLED preferred. Display resolution should not bottleneck sensor output.
- Refresh rate: 50Hz preferred for moving subjects. 25Hz adequate for stationary observation.
If any of these five specs are missing from the marketing materials, ask the manufacturer or find an independent review. Absent specifications are almost always absent because they're not competitive.
Frequently Asked Questions
What does NETD mean in thermal imaging? NETD stands for Noise Equivalent Temperature Difference — the smallest temperature change the sensor can reliably detect. Lower is better. ≤40mK is the target for serious outdoor use; it means the device can detect temperature differences of 0.04°C, which is necessary for clear imaging in warm ambient conditions.
Is 256×192 resolution good enough for wildlife observation? At distances under 150 meters for most subjects, yes — particularly with quality AI super-resolution processing. For observation at 200+ meters, or for applications requiring fine identification detail, 384×288 native resolution provides a meaningful improvement.
What lens size should I choose? For scanning-focused use (finding animals across a landscape), a 15mm lens with wide field of view. For extended-range observation of specific subjects you've already located, a 35mm lens with narrower, deeper field of view. Most first-time buyers benefit more from the wider 15mm lens.
Does 50Hz refresh rate matter for wildlife watching? For slow-moving subjects (deer browsing, raccoons foraging), 25Hz is adequate. For fast-moving subjects (running animals, birds, bats), 50Hz provides meaningfully smoother tracking. If fast-moving subjects are a regular observation target, 50Hz is worth prioritizing.
Why is the f/number missing from some thermal spec sheets? Manufacturers whose devices use slower apertures (f/1.2+) frequently omit the f/number because it's not a competitive specification for them. Missing aperture information should prompt investigation — contact the manufacturer or find independent test data.
