Unbreakable chain of intelligence
Protect the Integrity of Every Pixel
Optics, sensor, processor and AI, engineered as one system.
No downstream AI can recover details that aren’t captured. Sapient engineers every stage of the imaging chain — from capture to delivery — to the same rigorous standard, so everything the optics can see is there for the operator to act on.
Layer 1 · Optics
Aerial lens technology
Glass matched to mission altitude and ground resolution — wide-area to long standoff — resolving to the full 10K grid, and geometrically calibrated so each pixel maps to a known ray.
The optics layer →Layer 2 · Sensor
10K global-shutter capture
A 100 MP large-format square sensor at 10,000 × 10,000 pixels, all pixels exposed at once. The global shutter keeps every frame geometrically consistent from a moving platform — the basis for accurate detection and geolocation.
The sensor layer →Layer 3 · Compute
Every pixel processed onboard
Edge compute moves the full 10K stream over PCIe and runs inference on every pixel in real time, within available edge power and thermal budget — so only findings need the downlink.
The compute layer →Layer 4 · Software
Optimize, infer, deliver
Your AI model runs unchanged in an isolated container, fed tiles matched to its input, accelerated for the edge — and findings leave in open standards your systems already read.
The software layer →What engineering all four layers together buys you
- Direct georeferencingOptics + Compute
Geometric calibration, IMU, and GNSS combine into a camera pose per frame — pixels become target coordinates without a ground workflow.
- Environmental robustnessOptics + Compute
Temperature-compensated glass and robust enclosures hold performance across the altitude and temperature envelope — engineered for the field, not the bench.
- Open architecture · MOSASoftware + Compute
Run the stack on FORGE or on your own NVIDIA hardware — and a trusted, NDAA-compliant supply chain behind all of it.
01 · Optics
The Optics: The Image the Sensor Receives
A sensor can only record the image projected by the lens. The optics decide whether the sensor’s resolution is actually put to use, whether that holds up across the temperature, vibration, and altitude envelope of real missions, and — through careful calibration — whether a pixel can be turned into a precise, real-world location.
At a glance
- Precision glass, not plastic, holds sharpness edge-to-edge and across temperature.
- Resolving power is matched to the sensor, so no detail is lost at the lens.
- Geometrically calibrated — every pixel maps to a known angle, the basis for georeferencing.
Glass versus plastic
Lens elements are made from optical glass or from polymer (plastic). Glass is optically superior and far more stable: it expands very little and holds focus as the air cools with altitude and as the camera heats and pressure changes. Plastic is lighter and cheaper, but it shifts with temperature — moving the focus and softening the image just as the environment changes.
Resolving power
A 100 MP sensor only delivers 100 MP of real detail if the lens can render detail that fine. Every lens has a resolving limit: below a certain size, fine features blur together. Matching the lens’s resolving power to the sensor grid is what makes a high pixel count real rather than nominal.
Geometric calibration
Every real lens distorts the image. Geometric calibration measures that distortion and the exact internal geometry of the camera, so each pixel corresponds to a precisely known ray leaving the camera. Combined with the platform’s position and orientation, that is what turns a pixel into a coordinate on the ground: georeferencing, computed onboard rather than in a separate downstream workflow.
Aerial optics engineered to resolve the full 100 MP grid, temperature-compensated to hold focus across the operating envelope, with geometric calibration as the optical basis for onboard georeferencing. See the ROOK & ECHO sensors →
02 · Sensor
The Sensor: Turning Light Into Data
A digital image sensor is a grid of tiny light-collecting wells — the pixels. The size and number of those wells, and how they capture and read out light, set a hard ceiling on image quality. Information that is never captured cleanly cannot be recovered later, so the combination of optics and sensor sets the hard limit on what the AI model can detect.
At a glance
- A large square sensor can gather far more light, and full 12-bit RAW is kept end-to-end — so detail survives in deep shadow and bright highlight where 8-bit pipelines clip.
- An electronic global shutter freezes fast motion with no rolling-shutter skew and no moving parts to wear.
- Passive imaging: the sensor emits nothing, so nothing gives the platform away.
- Resolution and sensitivity are balanced for wide-area detection under real-world conditions, not just a pretty picture.
Sensor size and pixel size
A 35mm “full-frame” sensor measures 36 × 24 mm — many times larger than the physical area of the small sensors in phones, action cameras, and most video cameras. A larger sensor allows larger pixels, and a larger pixel is a bigger bucket: it can gather more photons before it fills up. More photons per pixel means a stronger signal relative to noise and a wider dynamic range — more usable data in deep shadows and bright highlights at the same time.
Resolution: more pixels, or bigger pixels
Resolution is a trade-off. On a sensor of fixed size, more megapixels means smaller pixels — each collecting less light. Keeping pixels large means fewer of them. A physically larger sensor is what lets you have both: a high pixel count with pixels that are still large enough for good data quality. What matters operationally is how many pixels land on the object — a measure called ground sample distance (GSD). Too few and a target is an unresolvable blob; enough and it becomes a recognizable shape. GSD has to be chosen to match the mission and the size of the objects that need to be detected and recognized.
Electronic global shutter
A sensor can be read two ways. A rolling shutter scans line by line, top to bottom, so on a moving platform image geometry skews and fast objects smear. A global shutter exposes every pixel at the same instant, recording one geometrically consistent frame. “Electronic” means it does this with no moving mechanical shutter — nothing to wear or fail — and with very short, precise exposures that freeze motion and reduce blur.
Passive imaging
A camera is a passive sensor: it only receives ambient light. Active sensors — radar (including SAR) or lidar — emit their own signal to see. That emission is itself detectable and can be jammed. A passive sensor transmits nothing to intercept, though it depends on there being light in the scene to begin with.
Color versus sensitivity
A color sensor has red, green, and blue filters over the pixels in a pattern called a Bayer array. Each pixel sees only part of the spectrum, and full color can be reconstructed by interpolation. A monochrome sensor has no color filters, so every pixel collects the full available light — giving more sensitivity, higher effective resolution, and reach into the near-infrared, which can help with seeing through smoke and haze. The trade is color information against raw light-gathering.
A 100 MP square sensor, available in RGB and monochrome / near-infrared variants, with an electronic global shutter. See the ROOK & ECHO sensors →
03 · Compute
The Compute: Why the AI Runs on the Aircraft
Sapient’s onboard processing is engineered to fit and run on a flying platform, so detection happens in real time and gives operators faster, better-informed decisions — rather than sending imagery to a ground station or the cloud and waiting for an answer.
At a glance
- The full 10K image data is transferred to the processor over an ultra-high-bandwidth PCIe bus — nothing dropped or compressed.
- Every pixel is processed onboard; only findings leave the aircraft, never raw video.
- Serious AI within the UAV’s size, weight, and power budget.
- Findings keep flowing through a radio-link blackout — nothing is lost, nothing stalls.
Moving the data: a dedicated internal path
A 10K sensor produces an enormous amount of image data every second — far more than typical camera connections can carry. PCI Express (PCIe) is an ultra-high-bandwidth data path normally used between components inside computers; here, it forms a dedicated internal link between the sensor and the processing module, so the entire full-resolution image data reaches the GPU without any bottlenecks. It is separate from the radio link to the ground, and it is what makes onboard, full-resolution processing possible in the first place.
Process onboard, not on the ground
10K image data is far too much to send over a radio link. Transmitting raw imagery is slow, floods the link, and stops entirely if the link drops. Processing on the aircraft means only the findings — a handful of small messages — ever need to be transmitted.
A data center’s job in a UAV power budget
Edge compute has to deliver serious AI performance in a package small and light enough to fly, and shed the heat that performance produces without active cooling wherever the airframe allows it. On a small UAV, available power is often the tightest constraint; on larger platforms it usually is not, since flight itself consumes far more power than compute ever will. What stays constant across platform sizes is size and weight — every gram and cubic centimeter added competes directly with fuel, payload, and range — with thermal management as the third factor: a sealed processing module still has to shed its own heat with no fan.
Keep everything; keep working
Fast onboard storage retains every raw frame for detailed review after the flight, while only findings are sent in real time. And because the processing is self-contained, the mission keeps working even if the communications link is jammed or lost — findings simply queue until the connection returns.
Many sensors, tightly synchronized
When fusing data from multiple sensors on the same platform — several cameras for a wider swath, or a mix of sensor modalities — Sapient precisely synchronizes capture timing across all sensors, so the data can be accurately combined.
A dedicated processing module based on the NVIDIA Orin NX, with a 2 TB NVMe SSD, enclosed in a compact, robust package that supports standard UAV interfaces. See the FORGE module →
04 · Software
The Software: From Frame to Finding
Hardware captures the image; software turns it into an answer. The enabling ideas here are less visible than glass and silicon, but they decide whether an AI model can run on the aircraft at all, how well it performs, and how quickly intel can be acted on.
At a glance
- Your AI model runs unchanged inside an isolated, encrypted container.
- Input is formatted exactly the way the model expects — scaled, tiled, calibrated.
- Fast enough to keep pace with the sensor in real time.
- Outputs speak the receiver’s language — Cursor-on-Target (CoT), straight into ATAK and C2.
Containers: a sealed box for the model
A container packages an application together with everything it needs to run, so it behaves identically wherever it runs and cannot reach anything outside its box — like a shipping container: standard on the outside, your cargo sealed inside. Each AI model runs in its own container; updating it means swapping the box rather than rebuilding the system, and because the box is sealed, the model and its data stay isolated.
Feeding the model what it expects
AI models are built for input images of a fixed size and a particular ground resolution, with an optimal trade-off between coverage and accuracy. A 10K frame is far larger than a model’s input size, so the image data is sliced into tiles that fit. With calibrated optics, camera pose, and a terrain model, the tiles are created and scaled so every tile has the optimal ground resolution.
Speaking the receiver’s language
A finding is only useful if the systems downstream can understand it. Sapient’s software transmits findings in the standard protocols and formats already used across command-and-control systems, so nothing new has to be developed or installed on the ground station to receive and use them.
Models run in isolated Docker containers, accelerated for the edge with TensorRT, and findings are delivered as Cursor-on-Target (CoT) into ATAK and existing C2. See the IGNITE:AI framework →
Designed together
Capabilities No Single Layer Could Deliver
The defining capabilities of the platform sit between the layers — they exist because optics, sensor, processor, and software are engineered as one system.
See Why the Engineering Matters
Explore sets out what these technologies buy you in the field — coverage, altitude, and a near-silent RF footprint.