Space is a rendering problem with unforgiving units. Objects move at kilometers per second, positions are good to the meter, the scene spans from a spacecraft's antenna to the 384,000 kilometers between Earth and the Moon, and the frame you compute in is never the frame you display in. We build the software that turns ephemerides, tracks, and sensor returns into something an operator can see and act on in real time.
AlphaPixel has helped build vision-based navigation for a (successful!) autonomous lunar lander flying with no GPS and no prior map, on-orbit vision navigation for space domain awareness, and 3D visualization for orbital and cislunar operations, all of it reaching the field through government and defense programs. This is where our computer vision, geospatial, and real-time 3D graphics work meet.
Space domain awareness
Catalogs of resident space objects that update continuously, conjunction and close-approach visualization, sensor-tasking and coverage displays, and the propagation underneath: SGP4/SDP4 from TLEs, or numerical propagation when two-line elements are not accurate enough. The hard part is drawing tens of thousands of objects moving at orbital velocity without the display stuttering or the picks going wrong, and doing it so an operator trusts what they see. When that picture carries military symbology, we generate it with our Nobori MIL-STD-2525 SDK.
Orbital mechanics and the precision problem
Earth-centered inertial, Earth-fixed, and geodetic frames; UTC, UT1, GPS, and TAI time; precession and nutation. Get any of them subtly wrong and the track drifts off the target. And at planetary scale, single-precision floating point falls apart. You get jitter and Z-fighting when a lander sits next to a mountain a third of a million kilometers from the coordinate origin. We render camera-relative with double-precision origins so the scene stays solid from orbit down to the surface.
Lunar and planetary terrain
Whole-body terrain built from real elevation and imagery (LRO/LOLA-class lunar data, planetary DEMs), correlated with our geospatial stack for landing-site selection, hazard and slope maps, and mission rehearsal. The terrain a descent camera sees in simulation is the terrain the operator reviews on the ground. We build these worlds with osgEarth, and increasingly with Rocky, its Vulkan-based successor, and we use both heavily for space-domain visualization. For physically accurate sensor and planetary rendering we also work with PANGU and SURRENDER, the tools that generate camera and sensor imagery for descent, landing, and rendezvous. We know the geodesy of Mars and the Moon.
How it connects
Descent and on-orbit navigation is a computer-vision problem. The terrain and coordinate systems come from geospatial. The rehearsal and operator displays are visual simulation built on OpenSceneGraph and Vulkan. Space work pulls the whole stack together, which is why one team that has built all of it beats four vendors who each own a single piece.
Who works on it
AlphaPixel is a US-owned small business, founded in 2004, with senior developers who each have 25-plus years in 3D, imaging, and simulation. DLA DD2345 / ITAR registered. We take closed-source defense work and open source alike, and we do not sell a product, so we have no reason to steer you toward one.
Frequently asked questions
What space software does AlphaPixel build?
We turn ephemerides, tracks, and sensor returns into something an operator can see and act on in real time. That covers space domain awareness displays, orbital and cislunar operations visualization, lunar-landing simulation, and on-orbit vision navigation, all reaching the field through government and defense programs.
What is space domain awareness, and can you build SDA visualization?
Space domain awareness is knowing where the resident space objects are and what they are about to do. We build continuously updating catalogs, conjunction and close-approach visualization, and sensor-tasking and coverage displays. The hard part is drawing tens of thousands of objects moving at orbital velocity without the display stuttering or the picks going wrong, and we do that so an operator trusts what they see.
How do you keep a scene precise across orbital and planetary scales?
Single-precision floating point falls apart when a lander sits next to a mountain a third of a million kilometers from the coordinate origin, producing jitter and Z-fighting. We render camera-relative with double-precision origins, so the scene stays solid from orbit down to the surface. We also handle the frame and time systems (ECI, ECEF, geodetic, UTC, UT1, GPS, TAI, precession, and nutation) that quietly drift a track off target when they are subtly wrong.
Do you support orbital propagation like SGP4/SDP4 and TLEs?
Yes. We propagate with SGP4/SDP4 from two-line elements, and switch to numerical propagation when TLEs are not accurate enough for the job. The propagation feeds the same real-time display the operator watches.
Can you do lunar or planetary terrain and landing-site work?
Yes. We build whole-body terrain from real elevation and imagery (LRO/LOLA-class lunar data and planetary DEMs), correlated with our geospatial stack for landing-site selection, hazard and slope maps, and mission rehearsal. For physically accurate sensor and planetary rendering we also work with PANGU and SURRENDER.
Do you do vision-based navigation for descent, landing, or on-orbit?
Yes. We helped on a project doing vision-based navigation for a successful autonomous lunar lander flying with no GPS and no prior map, and on-orbit vision navigation for space domain awareness. The terrain a descent camera sees in simulation is the terrain the operator reviews on the ground.
Can you work on classified or export-controlled space programs?
Yes. AlphaPixel is DLA DD2345 and ITAR registered and takes closed-source defense work and open source alike. We are used to the constraints that come with export-controlled and mission-critical software.
Why hire one small team instead of several vendors for a space program?
Space work pulls the whole stack together: descent and on-orbit navigation is computer vision, the terrain and coordinate systems are geospatial, and the operator displays are visual simulation. One team that has built all of it beats four vendors who each own a single piece. Start at our contact page.
Talk to us
Tell us what you’re building and where it’s stuck. We’ll tell you straight whether it’s something we can help with, and how we’d approach it.
Or see how we run consulting and software development engagements.
