More question? Contact us.
Thermal camera cores, such as those designed and integrated by Senssight, are versatile components used across a wide spectrum of high-stakes environments. Their primary function—converting infrared radiation into actionable visual data—enables them to serve several critical roles:
Industry 4.0 & Industrial Inspection: This is a major application where thermal cores monitor production lines and equipment in real-time. They are used for predictive maintenance, such as identifying overheating bearings, overloaded electrical cables, loose connections, or cooling system inefficiencies. By detecting these thermal anomalies before failure occurs, companies can avoid costly downtime and improve worker safety by conducting inspections remotely.
Surveillance & Security: Thermal cameras provide 24/7 visibility regardless of lighting conditions. They are essential for protecting property and people by detecting unauthorized intrusions in complete darkness, adverse weather, or obscured environments (e.g., fog or smoke) where standard optical cameras fail.
Drone Integration: Due to their small form factor and high-performance, thermal cores are integrated into UAVs (drones) to perform wide-area inspections, such as monitoring large-scale solar farms for faulty cells, inspecting power lines from the air, or assisting in search-and-rescue operations.
Fire & Rescue: Thermal imaging is a standard lifesaving tool for firefighters, allowing them to “see” through smoke to locate trapped victims, identify the seat of a fire, detect hidden hot spots within walls or ceilings, and monitor structural integrity to prevent collapses during an incident.
Thermography & Diagnostics: Beyond maintenance, thermal cores are used in building diagnostics to identify thermal bridges, air leaks, or moisture infiltration in insulation. In the medical or food industries, they can also be used for process control and quality assurance where non-contact, precise temperature measurement is required.
Essentially, whenever a professional needs to perceive the “invisible” heat energy emitted by an object or environment to gain situational awareness, perform non-invasive diagnostics, or ensure safety, a thermal camera core is the enabling technology.
To determine the specific distance at which you will be measuring or imaging targets, we have to look at how your operational range requirements dictate your choice of optics, field of view (FOV), and angular resolution. In thermal imaging, distance isn’t just a number—it dictates the entire physical configuration of the camera core.
We generally categorize imaging distances into three operational tiers:
1. Short-Range (Close-up to 50 meters)
Typical Applications: Industrial predictive maintenance, building diagnostics, PCB inspection, and indoor surveillance.
Optical Requirement: Wide-angle lenses (e.g., HFOV > 45°).
The Technical Trade-off: At close distances, you need a wide FOV to capture large equipment or rooms in a single frame. Because the target is close, even a wide FOV provides enough pixels on the target to achieve high-resolution imagery and accurate radiometric temperature measurements.
2. Medium-Range (50 meters to 500 meters)
Typical Applications: Perimeter security, drone-based solar farm or power line inspections, and standard search and rescue.
Optical Requirement: Medium/Standard lenses (e.g., HFOV between 15° and 45°).
The Technical Trade-off: This is the balancing act. The lens must be narrow enough to concentrate pixels onto a human or vehicle-sized target hundreds of meters away to satisfy DRI (Detection/Recognition/Identification) criteria, yet wide enough that a drone or security mast doesn’t lose situational awareness.
3. Long-Range (500 meters to several kilometers)
Typical Applications: Border control, coastal surveillance, counter-UAS (drone defense), and military reconnaissance.
Optical Requirement: Narrow telephoto or continuous zoom lenses (e.g., HFOV < 10°, often down to 1° or 2°).
The Technical Trade-off: At these distances, atmospheric interference (humidity, dust) becomes a major factor, and your angular resolution (measured in milliradians, mrad) must be incredibly tight. You need a long focal length lens to stretch your sensor’s pixels across a tiny, distant patch of space so that a target miles away still covers enough pixels to be identified.
The Crucial Metric: IFOV (Instantaneous Field of View)
When defining your distance, the metric that ties your resolution, pixel pitch, and optics together is IFOV (Angular Resolution). IFOV represents the exact angle of view covered by a single pixel on the sensor, calculated as:
If you know your target distance D and your lens’s IFOV, you can calculate the physical size of the smallest spot a single pixel can see (Spot Size = IFOV \ D).
For pure imaging (DRI): A human target at 1,000 meters needs to span at least 1.5 pixels wide just to be detected.
For temperature measurement (Radiometry): To get an accurate temperature reading, the target typically needs to cover a minimum of a 3×3 or 5×5 pixel grid to avoid “pixel bleeding” (slurring energy from the background). Therefore, a camera that can detect a human at 1 km cannot accurately measure their temperature at that same distance.
When identifying the temperature range your system needs to detect, you are defining the camera’s scene dynamic range. This is the window between the lowest and highest temperatures the camera sensor can accurately measure or image without “clipping” (saturating into pure white or bottoming out into pure black).
To configure a thermal core (like Senssight’s VGA or SXGA modules) for your specific application, we must look at two distinct temperature metrics:
1. Scene Temperature Range (What you are looking at)
This is the physical temperature of your target, typically broken down into standard industry tiers:
Standard / Ambient Range (-10°C to +120°C: This is the baseline configuration for standard Senssight cores. It is optimized for security, human/vehicle detection, night vision, and basic drone surveillance where targets remain around ambient environmental temperatures.
Industrial / Extended Range (-25°C to +250°C up to 650°C: Essential for industrial predictive maintenance, electrical inspection, and PCB testing. This allows the camera to detect hot electrical joints or mechanical friction without blinding the sensor.
High-Temperature Range (up to 1200°C or 2000°C): Required for specialized applications like kiln monitoring, furnaces, metallurgy, or advanced firefighting (where pinpointing the extreme seat of a fire is necessary).
2. Thermal Sensitivity vs. Measurement Range
Defining your target range requires balancing two technical capabilities: imaging (seeing contrast) and radiometry (measuring the exact temperature).
For Pure Imaging (Observation): If you only need to see an intruder or detect a hot spot, the exact temperature value doesn’t matter. What matters is the camera’s NETD (Noise Equivalent Temperature Difference). A sensor with an excellent NETD of <35mK (0.035°C), typical of Senssight cores, can differentiate between two objects that vary by a mere fraction of a degree, creating an incredibly sharp image even if the overall environment is cold.
For Radiometry (Measurement): If your application requires knowing the exact temperature of a target (e.g., “Is that bearing exactly 84°C?”), you need a radiometric factory calibration. Senssight cores, for instance, offer a typical radiometric accuracy of 2°C or 2% within their calibrated scene windows.
In the thermal imaging industry, determining how accurate your temperature measurements must be depends heavily on whether your application requires simple thermal anomaly detection (trending) or absolute precision (compliance and quality control).
When evaluating a thermal camera core, “accuracy” is actually split into two distinct performance metrics: absolute tolerance (accuracy) and repeatability.
1. Absolute Tolerance (Accuracy)
Absolute tolerance is the maximum allowable error between the camera’s displayed temperature and the target’s true thermodynamic temperature.
The Industry Standard: For high-quality radiometric thermal cores, the standard accuracy specification is +/-2°C or 2% (whichever is greater) of the reading in degrees Celsius.
What this means in practice: If you are measuring an object that is exactly 50°C, a standard +/-2°C tolerance means the camera is performing perfectly if it reads anywhere between 48°C and 52°C. However, if you are measuring a high-temperature industrial kiln at 500°C, the 2% rule applies, meaning the acceptable variance widens to 10°C (reading between 490°C and 510°C.
2. Repeatability (Precision)
Repeatability defines the camera’s ability to provide the exact same temperature reading under unchanged conditions over a period of time.
Why it matters: Even if a camera has an absolute tolerance of +/-2°C, you do not want the temperature readout on a stable object to constantly jump wildly between 48°C and 52°C every few seconds.
The Industry Standard: High-end cores generally maintain a repeatability of +/-1°C or 1%. This ensures that for automated predictive maintenance or quality control lines, the camera delivers a stable baseline, making it highly reliable for identifying sudden thermal spikes or gradual “trending” over weeks and months.
The Hidden Variable: Emissivity and Environment
It is crucial to note that a thermal core’s laboratory accuracy (+/-2°C is measured using a perfect laboratory instrument called a “blackbody.” In the real world, your measurement accuracy is highly dependent on software and environmental factors:
Emissivity (epsilon): Different materials emit infrared radiation differently. Highly reflective surfaces (like shiny copper or aluminum) have low emissivity and are notoriously difficult to measure accurately because they act like mirrors, reflecting the thermal signatures of the surrounding environment instead of showing their own heat.
Ambient Drift: Uncooled Long-Wave Infrared (LWIR) sensors are highly sensitive to their own internal temperature changes. To combat this, advanced cores use internal shutters and NUC (Non-Uniformity Correction) algorithms to frequently re-calibrate on the fly, maintaining measurement stability even if the ambient environment heats up or cools down.
When analyzing how fast a scene is changing—its temporal dynamics—you are directly determining the required frame rate (measured in Hertz or frames per second) of your thermal camera core. In the thermal imaging world, capturing fast-moving objects or transient thermal events requires a careful balance between frame rate and sensor exposure capabilities.
Frame Rate Tiers and Applications
Thermal camera cores typically fall into three primary frame rate categories, depending on the speed of the action they need to capture:
1. Standard Commercial Rate (9 Hz / <10 Hz)
The “Export-Controlled” Limitation: Because high-frame-rate thermal technology has military implications, international regulations (such as ITAR or dual-use export controls) strictly limit the export of thermal cameras operating above 9 Hz without extensive licensing.
When it works: 9 Hz is perfectly adequate for static or slow-moving scenes. This includes building diagnostics, stationary electrical panel inspections, environmental monitoring, or basic security surveillance where targets are walking.
The Drawback: If a target moves quickly across the frame, a 9 Hz camera will produce noticeable motion blur, choppy video, and can completely miss transient, split-second thermal spikes.
2. High Frame Rate (30 Hz to 60 Hz)
The Industry Standard for Motion: Cores operating at 30 Hz or 60 Hz (like standard Senssight configurations) deliver smooth, fluid video that matches the refresh rates of standard monitors.
When it works: This is the baseline required for drone integration (to compensate for aircraft vibration and forward flight speed), traffic monitoring, mobile robotic platforms, and standard manufacturing assembly lines.
Why it matters: At 60 Hz, the camera captures an image every 16.6 milliseconds. This drastically reduces motion blur, ensuring that a moving vehicle or a spinning mechanical component remains sharp enough for accurate DRI analysis or automated detection.
3. Ultra-High Frame Rate (>100 Hz to 1000+ Hz)
Specialized Scientific Research: Reserved for highly dynamic, transient thermal events.
When it works: Ballistics testing, automotive airbag deployment, fast-moving semiconductor testing, and advanced industrial welding analysis (such as laser or arc welding process control).
The Technical Challenge: Capturing hundreds of frames per second requires incredibly short exposure times. In thermal imaging, this means the sensor has less time to collect infrared photons for each frame, which can degrade the camera’s thermal sensitivity (NETD) unless specialized, cooled detector technology is used.
The Radiometric Catch: Frame Rate vs. Integration Time
If your application involves temperature measurement (radiometry) of a moving target, frame rate is only half the battle. You must also consider thermal time constant and integration time:
Sensor Integration Time: This is the “shutter speed” of the electronic sensor. If a hot component on a high-speed assembly line moves across the field of view faster than the pixel’s integration time, the heat signal will smear across multiple pixels. The camera will report a temperature that is lower than reality because the energy was diluted across space and time.
Transient Anomalies: In applications like laser welding, a defect (like a micro-crack or splatter) may heat up and cool down in a matter of milliseconds. A slow frame rate camera will completely miss the peak temperature of that transient event, logging only a averaged, lukewarm reading.
Summary Checklist
| Target Speed | Example Application | Recommended Frame Rate |
| Static / Human Walking | Building insulation, perimeter security | 9 Hz (Easy to export worldwide) |
| Vehicular / Airborne | Drone mapping, highway surveillance | 30 Hz – 60 Hz (Fluid tracking, low blur) |
| High-Speed Rotation | Spinning turbines, manufacturing lines | 60 Hz+ (Prevents thermal smearing) |
| Explosive / Transient | Welding physics, ballistics, air bags | >100 Hz (Specialized scientific cores) |
The physical environment where your thermal camera core will operate dictates its mechanical survivability, electrical protection, housing selection, and calibration stability. An uncooled thermal sensor (like those in Senssight modules) is a highly precise instrument; changes in the surrounding environment can directly alter its performance if the core is not properly ruggedized.
When designing or integrating a thermal camera system, you must engineer for five primary environmental threats:
1. Ambient Temperature Extremes
Thermal cores have two temperature specifications: Operating Temperature (the ambient conditions it can run in) and Storage Temperature.
The Challenge: Standard commercial cores typically operate from -20°C to +60°C. However, defense, aerospace, or heavy industrial environments (like steel mills or arctic surveillance) require extended ranges from -40°C to +80°C or higher.
The Impact on Calibration: Uncooled microbolometers are highly sensitive to their own internal temperature changes. Severe ambient shifts cause thermal drift, which degrades image clarity and ruins radiometric (temperature measurement) accuracy. High-quality cores utilize aggressive, real-time internal NUC (Non-Uniformity Correction) algorithms and automated mechanical shutters to constantly re-calibrate the sensor against its own internal temperature.
2. Dust, Ingress, and Particles
The Challenge: Operating in deserts, mining facilities, or cement factories introduces fine particulate matter that can scratch optics and clog internal electronics.
The Solution: The thermal core must be integrated into an appropriately rated enclosure—typically IP66 (dust-tight and protected against heavy seas) or IP67 (dust-tight and submersible).
The Optical Trap: Standard glass blocks infrared light. Therefore, the protective window of the camera housing cannot be regular glass; it must be made of specialized materials like Germanium (Ge) or Zinc Selenide (ZnSe) that allow Long-Wave Infrared (LWIR) wavelengths to pass through. These specialized windows often require hard-carbon coatings to resist scratching from windblown sand or dust.
3. Humidity, Moisture, and Corrosion
The Challenge: Tropical climates, marine environments, or industrial washdown zones introduce high humidity and salt fog.
The Solution: Moisture inside an enclosure will ruin electronics and cause internal fogging on the lens. Ruggedized housings are typically sealed and nitrogen-purged to eliminate internal moisture entirely. For marine settings, aluminum housings must be treated with anti-corrosive coatings (like passivated or marine-grade paints) to prevent galvanic corrosion.
4. Shock and Vibration
The Challenge: Cameras mounted on drones, military vehicles, industrial robotic arms, or pan-tilt security masts are subject to constant high-frequency vibration and sudden mechanical shocks.
The Solution: Standard commercial circuit boards and loose optical assemblies will fail under these conditions. Ruggedized thermal cores feature reinforced structural frames, potted electronics, and locked, ruggedized lens mounts (such as threaded M12 or custom bayonet interfaces) to ensure the lens does not drift out of focus during intense movement.
Summary Environmental Mapping
| Environment | Primary Risk | Mitigation Strategy |
| Outdoor Marine / Coastal | Salt corrosion, humidity, moisture | IP67 nitrogen-purged housing, marine-grade coating |
| Drone / UAV Flight | High-frequency vibration, wind cooling | Lightweight core, ruggedized lens mount, shock-absorbing gimbals |
| Heavy Industrial / Foundry | High ambient heat, fine abrasive dust | Protective cooling jacket (air or water), hard-carbon coated Germanium window |
| Arctic Surveillance | Extreme cold (-40°C) | Internal housing heaters, extended temperature component selection |
In thermal imaging, sensor resolution (pixel count) and pixel pitch are the two fundamental specifications that dictate the physical makeup of a microbolometer array. Together, they determine how much spatial detail a camera core can capture, directly impacting both the image sharpness and the overall distance performance (DRI) of the system.
Here is a breakdown of how these two metrics interact:
1. Sensor Resolution (Pixel Count)
Sensor resolution refers to the total number of independent temperature-sensing pixels on the microbolometer detector, organized in a horizontal and vertical grid.
Common Industry Standards:
QVGA (320×240 pixels): Entry-level resolution. Ideal for short-range applications, basic proximity detection, and budget-conscious integrations where fine detail isn’t critical.
VGA (640×480 pixels): The current industry sweet spot for professional applications (and a core standard for Senssight modules). It provides four times the pixel density of QVGA, offering a crisp balance of wide coverage and excellent target detail.
SXGA / XGA (1024×768 or 1280×1024 pixels): High-definition thermal imaging. Used for long-range surveillance, wide-area drone mapping, and highly precise scientific research where missing a single pixel of detail could mean failure.
The Resolution Advantage
A higher pixel count allows you to either capture a much wider field of view without losing spatial detail, or zoom in on incredibly distant targets while keeping enough pixels on the target to satisfy Recognition and Identification criteria.
2. Pixel Pitch
Pixel pitch is the physical distance measured from the center of one individual pixel to the center of the next adjacent pixel, typically quantified in micrometers (µm or microns).
The Evolution of Pitch: Historically, thermal sensors featured a 25µm or 17µm pixel pitch. Modern, high-performance thermal cores have shrunk this down to 12µm or even 8.5µm.
Why Smaller is Better: Shrunk pixel pitch means you can pack more pixels onto a physically smaller sensor chip.
The Optical and Financial Ripple Effect
Smaller pixel pitch yields massive system-level advantages:
Smaller Optics: A 12µm sensor requires a physically shorter focal length lens to achieve the exact same field of view and range performance as an older 17µm sensor. Because Germanium lenses are highly expensive, shrinking the pixel pitch drastically reduces the size, weight, and cost (SWaP-C) of the entire camera lens assembly.
Higher Spatial Resolution (IFOV): By placing pixels closer together, the angular resolution improves. Each pixel samples a smaller, tighter point in space, minimizing pixel blur and allowing you to resolve incredibly small features on a target.
The Synergy: Spatial Detail vs. Sensitivity
While a higher resolution and smaller pixel pitch provide superior spatial detail, there is an engineering trade-off regarding thermal sensitivity (NETD):
NETD (Noise Equivalent Temperature Difference) is the primary metric used to quantify a thermal camera core’s thermal sensitivity. It defines the minimum temperature difference that the camera sensor can reliably detect against its own internal electronic noise.
In simpler terms, NETD tells you how good the camera is at distinguishing tiny variations in heat. It is typically measured in millikelvins (mK) or thousandths of a degree Celsius (°C).
How NETD Works
When a thermal camera looks at a scene with uniform temperature, the sensor still generates a tiny amount of inherent electronic noise. If the temperature difference between two objects in the scene is smaller than this internal noise, the camera cannot tell them apart, resulting in a grainy, “noisy” image.
The NETD specification marks the exact threshold where the thermal signal strength equals the camera’s internal noise (Signal-to-Noise Ratio = 1).
Real-World Impact: What Different NETD Ratings Mean
When evaluating a thermal core (like the high-performance Vanadium Oxide microbolometers used by Senssight), the NETD rating directly dictates image clarity and performance in challenging environments:
Excellent Sensitivity (<35mK to <40mK): This is the professional standard for high-end cores. At less than 0.035°C of sensitivity, the camera can resolve incredibly fine thermal details. Even on a cold, rainy day, or in a completely washed-out gray landscape, the camera will produce a sharp, high-contrast image, easily distinguishing footprints, minor insulation gaps, or subtle human thermal signatures.
Standard Sensitivity (<50mK): A solid industry baseline. Perfectly adequate for standard industrial predictive maintenance, automated manufacturing lines, and medium-range surveillance where targets (like running machinery or people) have a distinct thermal contrast against the background.
Low Sensitivity (>60mK to 100mK): Found in low-cost, entry-level, or highly legacy hardware. The resulting image appears noticeably grainy or “blurry” in low-contrast conditions. These sensors require a massive temperature difference between the target and the background to produce a clear outline.
Why NETD Matters for Your Application
An exceptional NETD is critical for two major operational reasons:
1. Low-Contrast Environments
In severe weather (thick fog, heavy rain, or snow), thermal energy is scattered and absorbed by the atmosphere. The temperature differences reaching the camera lens become heavily muted. A core with a low NETD (<35mK) can slice through this environmental degradation to extract a usable image, whereas a lower-quality sensor will show nothing but uniform gray static.
2. Advanced Diagnostic Analytics
In applications like building diagnostics, fluid level detection in tanks, or early-stage PCB inspection, the thermal anomalies you are searching for are often incredibly faint—sometimes a fraction of a degree. A highly sensitive core ensures these subtle gradients are captured and visualized before they escalate into a failure.
Selecting the appropriate spectral band is a fundamental design choice. For most thermal cameras—including standard Senssight modules—the system operates in the LWIR (Long-Wave Infrared) band, which spans wavelengths from 8 to 14 µm.
When building or integrating a thermal camera system using a core like Senssight’s, the lens selection is just as critical as the sensor itself. In the infrared spectrum, optical design is bound by rigid physical constraints regarding how Field of View (FOV), Focal Length, and specialized Lens Materials interact.
1. The Interplay of FOV and Focal Length
The Field of View (FOV)—the angular width of the scene the camera can see—is not a fixed property of the sensor. It is determined mathematically by the relationship between the sensor’s physical size (dictated by resolution and pixel pitch) and the lens’s focal length.
Because of this equation, focal length and FOV share an inverse relationship:
Short Focal Length (e.g., 4mm to 9mm): Yields a Wide FOV (e.g., 50° to 90°). This is ideal for close-range situational awareness, indoor security, and capturing massive industrial machinery in a single frame.
Long Focal Length (e.g., 35mm, 50mm, or 100+mm): Yields a Narrow FOV (e.g., <15°). This acts like a telescope, focusing all the sensor’s pixels onto a tiny patch of space far away, maximizing your long-range DRI (Detection, Recognition, Identification) performance.
2. Lens Materials: Why Regular Glass Won’t Work
Standard optical glass is completely opaque to Long-Wave Infrared (LWIR) radiation (8-14µm); it absorbs the heat energy instead of letting it pass through. Because of this, thermal lenses must be manufactured from exotic, specialized crystalline materials:
Germanium (Ge): The gold standard for LWIR imaging. Germanium has a very high refractive index, allowing for compact lens designs with exceptional optical clarity. However, it is a scarce element, making Germanium lenses the single most expensive component of a thermal camera system. It is also prone to thermal runaway (becoming opaque if heated past 100°C), requiring special considerations in high-heat industrial foundries.
Chalcogenide Glass: A synthetic glass alternative composed of elements like selenium, sulfur, and tellurium. It can be molded rather than diamond-turned, making it a highly cost-effective choice for high-volume commercial integrations, drone cores, or automotive night vision.
Zinc Selenide (ZnSe) / Zinc Sulfide (ZnS): Often used in multi-spectral systems or specialized industrial windows because they can transmit both visible light and infrared wavelengths simultaneously.
To protect these sensitive elements from scratching or environmental degradation, manufacturers apply rugged Hard Carbon or Anti-Reflective (AR) coatings to the outermost lens surface.
3. Interchangeable Lenses vs. Fixed Optics
Depending on how adaptable your application needs to be, thermal cores generally offer two optical mounting philosophies:
Fixed/Integrated Lenses
Many thermal camera cores come factory-sealed with a fixed focal length lens. This is typical for drone payloads, rugged outdoor security cameras, and compact handheld thermography tools.
Pros: Optimal factory calibration, better environmental sealing (IP67), lighter weight, and lower cost.
Cons: Zero optical flexibility; if your deployment distance changes, you must replace the entire unit.
Interchangeable Lenses (Threaded / Bayonet Mounts)
High-end professional and scientific thermal cores feature interchangeable lens mounts (often utilizing standard M12 x 0.5 threads, C-mounts, or proprietary bayonet locking mechanisms).
Pros: Ultimate versatility. You can mount a wide-angle lens for indoor factory monitoring on Monday, and swap it for a telephoto lens for outdoor perimeter security on Tuesday.
The Radiometric Challenge: If you swap lenses on a radiometric (temperature-measuring) thermal camera, you will lose measurement accuracy. Because every individual piece of Germanium glass transmits infrared energy slightly differently, a radiometric core must be factory-calibrated to the specific lens attached to it. Premium systems solve this by storing multiple “lens calibration profiles” directly in the camera’s firmware, allowing users to swap lenses and select the corresponding profile on the fly.
When deciding whether a thermal camera core should feature fixed, manual, or autofocus, the choice directly hinges on how dynamic your target distance is and whether you are performing pure imaging or absolute temperature measurement.
Here is how the three focus types compare across industrial and security applications:
1. Fixed Focus (Athermalized Lenses)
With fixed focus, the lens is factory-set to a specific distance matrix (often from a few meters out to infinity) and locked into place.
How it handles environmental shifts: Because temperature changes cause Germanium glass to expand or contract—which shifts the focus point—fixed lenses utilize athermalization. This mechanical design pairs the glass with housing materials that expand inversely, keeping the sensor in focus across wide temperature swings (e.g., -40°C to +80°C).
Best Used For: Proximity security, drone payloads (where weight and power must be minimized), helmet-mounted cameras, and highly standardized close-up assembly line monitoring.
2. Manual Adjustable Focus
Manual focus allows an operator to physically rotate the lens barrel, or use a low-profile motorized command via software, to pin-point a specific plane of clarity.
The Radiometric Rule: If you are using a camera for radiometry (measuring temperature), manual focus is highly reliable. Since out-of-focus thermal images cause heat energy to “bleed” across pixels—resulting in falsely low temperature readings—manually locking the focus ensures stable, uncompromised data during long inspection routines.
Best Used For: Stationary automated industrial inspection, electrical cabinet monitoring, lab-bench R&D, and fixed perimeter security masts where the target distance never changes.
3. Continuous Autofocus
Autofocus systems use internal digital signal processing (DSP) to analyze image contrast on the fly, actively driving a miniature internal motor to keep the scene perfectly sharp.
The Integration Challenge: Germanium is an incredibly dense, heavy material. Moving a large thermal telephoto lens rapidly requires high-torque stepper motors and robust gearboxes, which increases the core’s size, weight, and power consumption (SWaP). Furthermore, the autofocus algorithm must be sophisticated enough to ignore transient clutter (like a bird flying by or a burst of exhaust steam) so the camera doesn’t constantly “hunt” and blur the background.
Best Used For: Pan-Tilt-Zoom (PTZ) long-range border surveillance, tracking fast-moving targets, robotic vision platforms, and mobile handheld devices where an inspector is constantly changing targets.
Selection Matrix
To help you decide which configuration your Senssight core integration requires, consider this quick roadmap:
| Operational Scenario | Recommended Focus Type | Key Benefit |
| Fixed distance, constant environment | Fixed Focus | Lowest cost, highly rugged, zero moving parts to fail. |
| Fixed distance, precise temperature measurement | Manual / Motorized Lock | Prevents radiometric pixel bleeding; ensures absolute temperature accuracy. |
| Constantly shifting target distance or moving platform | Continuous Autofocus | Eliminates operator intervention; maintains critical tracking and DRI ranges automatically. |
The Minimum Focus Distance (MFD)—or close-focus capability—defines the closest physical proximity a target can be to the thermal lens while remaining sharply in focus. For applications like circuit board (PCB) inspection, medical research, or close-up industrial quality control, the MFD is a vital specification. If you position an object closer than the lens’s design threshold, the infrared radiation scatters before hitting the microbolometer sensor array, introducing severe geometric blur and corrupting radiometric temperature accuracy.
Optical Drivers of Close-Focus Performance
The physical limit of your camera core’s minimum focus distance is dictated by three primary factors:
1. Lens Focal Length and FOV
There is an engineering rule of thumb in infrared optics: the shorter the focal length (wider the lens), the shorter the minimum focus distance.
Telephoto Lenses (e.g., 24mm to 50mm): Long lenses compress the depth of field. A narrow 24mm or 35mm lens designed for mid-range tracking may have a native minimum focus distance of 1 meter to 3 meters. Trying to use this setup to inspect an electronics bench will yield nothing but a blurry, unusable thermal smear.
2. Lens Mount Mechanical Travel
To focus on an object that is incredibly close, the entire Germanium lens assembly must physically move forward (away from the sensor chip).
Manual Threaded Mounts (M12): Cores that use adjustable threaded lens barrels give the integrator manual flexibility. You can unscrew the lens slightly to extend its travel, which effectively pulls the focus plane much closer to the camera face.
Fixed Athermalized Lenses: If a core features a factory-sealed, fixed lens engineered for outdoor drone use, it is structurally locked to focus at long distances to infinity. It cannot move forward, completely ruling it out for close-range microscopic work.
Elevating Close-Focus: Macro Lenses vs. Extension Rings
If your application demands extreme close-up detail—such as identifying a micro-fracture on an integrated circuit or measuring the temperature of an individual resistor—a standard thermal core configuration will fall short. Integrators use two specialized methods to achieve Macro-level imaging:
Macro Lens Attachments: A secondary Germanium element is placed over the existing lens. It works like a magnifying glass, bending the incoming infrared rays sharply to create an ultra-short focus distance (often 3cm to 5cm) while providing microscopic spatial resolution (often down to tens of micrometers per pixel).
Extension Rings (Lens Spacers): These are simple mechanical rings inserted between the camera core body and the lens mount. By adding a physical 1mm or 2mm buffer space, the lens is pushed further away from the sensor. This shifts the entire focus range forward, trading away infinity focus to gain extreme close-up macro capabilities.
For Senssight’s thermal camera cores, the Instantaneous Field of View (IFOV)—or the angular size of a single pixel’s coverage on the target—is a critical specification for determining spatial resolution at a given distance.
How to Calculate IFOV
The IFOV is typically calculated using the formula:
IFOV (mrad) = (Pixel Pitch) / (Focal Length)
For example:
How to Use IFOV to Determine Spatial Resolution
To find the size of one pixel’s coverage on the target (in meters or millimeters), use:
Pixel Coverage (m) = IFOV (rad) × Distance (m)
Example:
This means each pixel covers 6.3 cm of the target at 100 meters.
Why IFOV Matters
Yes. All Senssight camera cores support radiometric mode, meaning every pixel in the image returns a calibrated temperature value — not just a visual thermal representation. This radiometric capability is included as standard and handled through our SDK, which provides full access to the raw temperature data stream for each frame.
Our in-house radiometric calibration process ensures a thermographic accuracy of ±2°C over an ambient temperature range of 0°C to 60°C and a scene temperature range of -25°C to 250°C. Each core is calibrated according to three profiles — standard, extended and precise — so you can select the level of accuracy that matches your application requirements.
This makes Senssight cores well suited for any application where temperature measurement matters: predictive maintenance, industrial inspection, fever screening, agricultural stress detection, and more.
Senssight camera cores use a multi-layer calibration approach designed to deliver consistent thermal image quality and radiometric accuracy across the full operating temperature range — without interrupting the video stream.
Factory calibration is performed in-house at Senssight. Every core could be radiometrically calibrated across a controlled range of ambient and scene temperatures, producing the correction tables that underpin the ±2°C thermographic accuracy.
Shutterless operation is the primary mechanism used to compensate for the natural drift of the infrared sensor over time and temperature. Rather than periodically blocking the optical path with a mechanical shutter to perform a non-uniformity correction (NUC) — which interrupts the video stream and can be a critical issue in tracking or detection applications — our shutterless calibration algorithms continuously correct for pixel-level drift in the background. The result is a stable, uninterrupted thermal video flow at all times.
Shutter-based offset correction remains available on cores equipped with a built-in mechanical shutter (Ernest and Niels families). When operating conditions require it — for instance after a significant sun exposure — a shutter-triggered NUC can be performed to reset the baseline and restore optimal image uniformity. This can be triggered automatically or manually via the SDK, giving integrators full control over the correction strategy.
In short: shutterless calibration handles the continuous drift, radiometric calibration underpins temperature measurement, and the optional shutter provides a hard reset when the environment demands it.
Senssight thermal camera cores with radiometric mode typically offer an absolute accuracy of ±2°C or ±2% of the reading (whichever is greater) and repeatability of ±1°C or ±1% under controlled conditions. These specs depend on proper calibration, emissivity settings, distance, and operating within the camera’s designed temperature range.
Yes, Senssight’s radiometric thermal camera cores allow you to adjust critical parameters for accurate thermography, including:
- Emissivity (0.01–1.0) to match the material being measured.
For radiometric thermal cameras, temperature stability (or thermal drift) refers to how consistently the camera maintains its calibration and accuracy over time and under varying ambient conditions. This is critical for applications requiring long-term reliability, such as industrial monitoring, predictive maintenance, or scientific research.
To maximize radiometry performance, Niels should be placed as far as possible from heat sources (e.g., processors or other high-power dissipating devices). Integration and fixation to the customer’s chassis should be done using the mounting holes on each side. Thermal dissipation of Niels should be considered, as reduced housing temperature generally leads to better overall performance and reliability.
The maximum frame rate options vary by product family depending on the sensor’s pixel count:
Eileen 3 (QVGA – 320×240): operates at 60Hz by default.
Ernest 6 (VGA – 640×480): Operates at a native high-speed default of 60 Hz, with an optional 30 Hz configuration choice.
Niels 12 (SXGA – 1280×1024): Operates at a maximum native speed of 60 Hz by default (with an optional 30Hz configuration).
Yes. The Senssight platform includes explicit software and hardware-level windowing management natively built into its technical architecture.
The standalone Senssight SDK provides full library support to configure a customized Region of Interest (ROI) or sub-window on the sensor array. By utilizing this windowing function to reduce the active readout resolution (e.g., cropping down the frame area), the system can streamline processing overhead, reduce active USB/MIPI data pipeline bandwidth constraints, and capture targeted regions at significantly enhanced readout speeds.
This layout allows you to dynamically trade spatial coverage for maximum frame rate performance depending on the strict frame-rate demands of your tracking, industrial, or drone-based system.
Across optimized host processing platforms, the system maintains a streamlined real-time data pipeline. Tested on an embedded platform (such as the NVIDIA Jetson Orin NX), the typical hardware-to-software latency between the end of the sensor array integration loop and raw frame availability inside the SDK callback environment (getRawFrame) is less than a single frame. For a core running at 60 Hz, this guarantees a predictable latency of less than 16.67 ms, assuming the host platform has the required processing power to ingest UVC data at full speed without retries.
Based on testing, the performance of the Raspberry Pi (both Pi 5 and Pi 4B) appears insufficient for video capture with the Ernest 6 or Niels 12, resulting in limited FPS and sped-up video. This occurs with both Senssight’s provided SDK and custom software built on it. Video capture works flawlessly at 60 FPS with the Eileen 3 due to its lower resolution. It is suggested that a faster processing unit might be needed for smooth video.
Senssight thermal camera cores natively utilize USB interfaces. The Eileen 3 (QVGA) and Ernest 6 (VGA) cores feature USB 2.0 Hi-Speed connectivity, while the high-resolution Niels 12 (SXGA) core uses a USB 3.1 Super-Speed interface.
The MIPI interface is scheduled for 2026.
Yes, the cameras appear as a standard video device. They use the USB UVC (USB Video Class) protocol for seamless image transfer alongside the USB CDC protocol for system configuration, eliminating the need for complex custom drivers.
The comprehensive Senssight SDK is supplied across all SeeCube products and includes necessary embedded calibration data. It supports widespread architectures and cross-platform deployments:
Windows x86-64
Linux Ubuntu/Debian (x86_64 or arm64)
Raspberry Pi OS
Nvidia Jetpack 6.x and 7.x
The pipeline streams uncompressed Y16 Big Endian raw data, a true 16-bit format that preserves the essential underlying radiometric metadata for temperature analysis.
The Senssight SDK provides powerful integrated enhancement features directly within its functional library, including native device handling, bad pixel removal, Automated Gain Control (AGC), histogram equalization, shuttered and shutterless mode management, and image colorization.
The SDK comes fully equipped with a wide variety of functional developer code examples demonstrating how to easily connect, configure, and record both raw images and videos.
The camera cores are entirely USB powered via a 5V supply voltage, meaning they do not require any external power bricks or specialized rails. Power draw is highly optimized: Eileen 3 consumes less than 400mW, Ernest 6 consumes less than 650mW, and Niels 12 requires 1.7W.
The hardware incorporates physical, space-saving layout connectors to match compact integration needs: Eileen 3: Utilizes a right-angle USB micro connector. Ernest 6: Offers versatile connector configurations including a USB Flex cable option, straight or right-angle USB-C (Male or Female), and a right-angle USB micro port. Niels 12: Implements a rugged USB3 micro-B layout that features 2x M2 structural holes to accommodate securely screwed USB3 cables.
The SDK supports Windows x86-64, Linux Ubuntu/Debian (x86_64 or arm64), and Raspberry Pi OS. Android 12+ is also supported, depending on chosen hardware compatibility and performance. For Niels 12 specifically, the SDK is available for Nvidia Jetson, Android, and Windows (coming soon).
The SDK includes functions for device handling and enumeration, bad pixels removal, Automated Gain Control (AGC) and histogram equalization, shuttered and shutterless operation, and colorization. Various examples are provided for connecting, configuring, and recording images and videos.
The thermal camera cores feature an ultra-lightweight, compact structural architecture optimized for extreme size, weight, and power (SWaP) constraints:
Eileen 3 (QVGA): Forms an ultra-small profile with a standard chassis weight of 15g (regardless of the lens).
Ernest 6 (VGA): Features a best-in-class tiny footprint of only 21×21 mm with a lightweight core chassis weighing 20g (or up to 23g depending on specific electrical connector configuration choices).
Niels 12 (SXGA): Serves as the high-resolution, small-form-factor core in the family, with a bare camera body weight of 60g to 63g.
The standalone thermal camera cores provide a highly resilient industrial/military operational base, supporting structural operating temperature range extending from -35°C to +75°C across all product tiers. For custom setups requiring specialized exterior deployments, please contact us.
Senssight camera cores are thoroughly proven for rugged, high-vibration applications such as industrial production plants and aerospace drone payloads. Rigorous environmental testing demonstrates full resilience under intense physical stresses without any interruption to the imaging pipeline:
Vibrations: Validated to full compliance under standard military mechanical environments via MIL-STD-883 H (Method 2026) testing.
Mechanical Shocks: Officially passed structural impact qualifications meeting both MIL-STD-810 G (Method 516.6) and MIL-STD-883 G (Method 2002.5).
Throughout and following these military-grade testing routines, internal shutter operations remain perfectly uncompromised and the raw thermal image streams stay fully uninterrupted.
The architecture varies flexibly depending on the core resolution family to accommodate different operational requirements:
Eileen 3 (QVGA): Features a completely shutterless hardware architecture by default (no physical mechanical shutter built into the core framework). Non-Uniformity Correction (NUC) is entirely handled via advanced software options.
Ernest 6 (VGA) & Niels 12 (SXGA): Designed with an integrated physical mechanical shutter directly onboard the chassis to facilitate automated standard NUC runtime recalibrations. However, these cores can be dynamically operated in a true shutterless environment by selecting the premium factory processing upgrades (_SL+ or _R1), which rely on pre-loaded software matrix corrections instead of mechanical shutter triggers.
Upon system startup, the uncooled internal sensor arrays exhibit a gradual internal temperature stabilization loop. Initial operational temperatures read near local ambient limits, subsequently warming to +10°C to +15°C over the ambient environment over the first several minutes of initialization before entering a steady thermal dissipation plateau. To guarantee the most stable radiometric precision across long-duration use-cases, the system interface provides embedded factory calibration tables that dynamically compensate for drifts across the sensor’s native microbolometer operational bounds.
Warranty period : 1 year
Senssight provides standard professional support alongside its core product deliveries, though it expressly notes in its documentation that it makes no warranty, representation, or guarantee regarding the specific suitability of its products for any particular purpose beyond standard factory operational specifications.
Advanced factory calibration is integrated directly into the product delivery based on the processing option chosen. Customers can select Native Shutterless (_SL), Enhanced Shutterless (_SL+), or fully calibrated Radiometric Shutterless (_R1) configurations where every individual image pixel is factory-calibrated to deliver absolute temperatures.
Built upon industry-proven uncooled microbolometer thermal imagers supplied directly by Lynred (such as the ATI320 and ATTO1280 sensors), the camera cores are engineered for extreme stability and high-end industrial or military applications.
System upgradeability and backward compatibility are maintained through a robust architecture; the software pipeline is entirely managed via the standalone Senssight SDK. The SDK contains embedded calibration data, provides comprehensive device handling, and explicitly supports continuous cross-platform deployment across Windows x86-64, Linux Ubuntu/Debian, Raspberry Pi OS, …
Senssight maintains a strong reputation for engineering quality and prompt technical responsiveness, operating directly out of France’s tech hub in Meylan and utilizing international distributors or sales agents. The documentation is highly detailed, providing clear mechanical constraints, precise 2D/3D step files, and extensive code integration examples within the SDK to streamline developer onboarding.
Pricing is structure-dependent on configuration and volume.
Generally, longer focal lengths are more expensive than shorter ones. As for processing options, the radiometric shutterless _R1 option is more expensive than the _SL+ (shutterless) option, which is more expensive than the _N (shutter) option.
Senssight offers processing upgrades from default _N (shuttered mode) including Native Shutterless (_SL), Enhanced Shutterless (_SL+), and Shutterless Radiometric (_R1) calibrations.
Seecube SDK is provided free of charge.
The thermal camera cores maximize return via industry-leading SWaP (Size, Weight, and Power) performance, an all-inclusive integration SDK, and unrivaled radiometric precision to minimize time-to-market.
Overpayment is prevented by selecting from an adapted product tier, scaling from the ultra-affordable Eileen 3 (QVGA) up to the high-resolution Ernest 6 (VGA) and Niels 12 (SXGA) cores.
All Senssight camera families are fully RoHS compliant.
Designed for harsh military and industrial application, the camera cores stay fully operational after undergoing rigorous military environmental testing, meeting MIL-STD-883H vibration standards and MIL-STD-810G shock requirements.
Niels camera core has passed vibration tests (MIL-STD 883 H, method 2026) and shock tests (MIL-STD 810 G, method 516.6 and MIL-STD 883 G, method 2002.5) without any damage. Shutter behavior did not change during tests or after tests. Image stream has not been interrupted during tests and after tests. Niels camera core stays fully operational after tests.
Ernest camera core has passed vibration tests (MIL-STD 883 H, method 2026) and shock tests (MIL-STD 810 G, method 516.6 and MIL-STD 883 G, method 2002.5) without any damage. Shutter behavior did not change during tests or after tests. Image stream has not been interrupted during tests and after tests. Ernest camera core stays fully operational after tests.
Eileen camera core has passed vibration tests (MIL-STD 883 H, method 2026) and shock tests (MIL-STD 810 G, method 516.6 and MIL-STD 883 G, method 2002.5) without any damage. Image stream has not been interrupted during tests and after tests. Eileen camera core stays fully operational after tests.
Because they utilize high-performance thermal imagers, these goods are classified as dual-use items subject to strict export control regulations outside the EEC.
Here is the trade regulation information for all thermal camera core over 9Hz:
Each thermal camera core has mounting holes on its outer surfaces. Please refer to the product datasheet for more detailed information.
Diverse standard lens options are tailored to system coverage requirements, spanning from a wide 86° HFOV down to a narrow 8° HFOV.
Please check out the Field of View simulator in the Resources tab on the website.
Thermal data is output as Y16 Big Endian raw frame data and handled through the Senssight SDK via standard USB UVC protocols.
Systems maintain streamlined pipelines utilizing USB 2.0 Hi-Speed or USB 3.1 Super-Speed interfaces, with MIPI integration launching in late 2026.
Cores run on a 5V supply with low power demands (<400mW for Eileen 3, <650mW for Ernest 6, and 1.7W for Niels 12). To maintain the optimal radiometry temperature range (10°C to 60°C), systems must be mounted away from external heat sources.
Units are factory calibrated based on the chosen processing grade, and live video integration or sensor accuracy can be verified using the calibrated data embedded in the Senssight SDK.
Senssight offers three main product families: Eileen 3 (QVGA 320×240 pixels), Ernest 6 (VGA 640×480 pixels), and Niels 12 (SXGA 1280×1024 pixels).
A new family a.k.a Ernest 10 (XGA 1024×960 pixels) will be available by the end of 2026.
By default, Senssight does not supply thermal camera cores without optics, especially the Eileen 3 version as it incorporates a Lynred ATI320. The main reason is that options like shutterless and thermography require calibration with the optics. Not all lenses are mechanically compatible with Senssight modules (back focal length, mount diameter, etc.).
Senssight typically does not supply thermal camera cores without optics because options like shutterless and thermography require calibration with the optics. For an Ernest 6 camera core without optics, visibility on potential volumes (one-off or future volumes) is needed to assess the possibility.
More question ?
