Orbital Computing for Earth Observation Satellites Versus High-Speed Optical Downlinks

Key Takeaways

• Optical downlinks reduce data pressure, but they do not remove every case for on-orbit computing.
• On-orbit processing fits triage, alerts, compression, autonomy, and low-latency users.
• Ground processing remains better for calibration, archives, science products, and broad reuse.

Orbital Computing for Earth Observation Satellites Has a Narrower Job Than It First Appears

NASA defines data latency as the elapsed time between instrument acquisition and public data availability, and that definition frames the real question behind orbital computing for Earth observation satellites. The question is less about whether a satellite can process imagery in space and more about which decisions should be made before raw or partly processed data reaches the ground. Earth observation (EO) satellites already create data products through multiple processing levels, from raw instrument data to higher-level products that convert measurements into geophysical information such as temperature, land cover, moisture, or deformation. NASA Earthdata treats latency as a full chain issue, not a spacecraft-only issue, because acquisition, downlink, ingest, processing, quality control, distribution, and user access all add time.

Orbital computing, also called on-orbit computing or edge computing in space, shifts some data handling from ground systems to the satellite. That can include compression, cloud screening, change detection, event recognition, data prioritization, sensor retasking, or early product generation. The concept has gained attention because EO instruments can collect more data than many missions can transmit in a timely way, and because customers increasingly want alerts rather than files. The European Space Agency’s Φsat-1 demonstrated artificial intelligence for EO by filtering out less useful imagery before transmission, and Φsat-2, launched on 16 August 2024, extends that idea with a multispectral imager and onboard artificial intelligence framework.

High-speed optical downlinks challenge the need for heavy onboard processing because they raise the amount of data that can move from orbit to Earth. NASA states that laser communications can transmit 10 to 100 times more data than comparable radio frequency systems, and NASA’s TBIRD technology demonstration completed a mission after breaking records for high-rate optical downlinks from space.

The result is not a simple replacement of onboard processing by optical downlinks. Faster downlinks reduce the need to discard data in orbit, but they do not eliminate the value of deciding which data matters first. Weather over optical ground stations, limited contact windows, pointing requirements, onboard storage, ground network scheduling, data rights, customer latency requirements, and defense and security use cases can still favor onboard triage. At the same time, many science and commercial applications need carefully calibrated, traceable, reusable data products that ground infrastructure handles better than a small spacecraft computer.

Optical Downlinks Change the Data Bottleneck

Radio frequency downlinks created a long-running constraint for EO missions. A satellite could collect imagery during each orbit, store it onboard, and wait for access to a compatible ground station. Higher-resolution sensors, wider swaths, hyperspectral instruments, synthetic aperture radar, and video imaging all increased the size of the data stream. When the downlink pipe stayed narrow, operators had three options: collect less data, add more ground stations, or process data onboard so that fewer bits needed transmission.

Optical communications change that trade. Laser links use narrow beams of light rather than broader radio beams, allowing much higher data rates under the right conditions. NASA’s Laser Communications Relay Demonstration tested relay concepts for moving data between space and Earth, and NASA’s ILLUMA-T payload later completed the agency’s first two-way, end-to-end laser relay system with LCRD before the ILLUMA-T mission ended in June 2024. Those demonstrations were not EO missions, but they matter for EO architecture because the same communications principle can move large observation files faster.

The European Data Relay System, branded as the SpaceDataHighway, shows the same logic in an EO setting. The system relays data from low Earth orbiting satellites through geostationary communications satellites to Europe, reducing dependence on direct ground-station visibility. Copernicus reported that EDRS laser terminals work at up to 1.8 Gbit/s and can relay large volumes of Sentinel data, including data with value for maritime surveillance and emergency response. ESA describes EDRS as a way to transmit large quantities of EO data in near-real time.

Optical downlinks do not make every bottleneck disappear. Direct-to-Earth optical links need suitable weather at the receiving site because clouds can block or degrade the path. Narrow beams demand accurate pointing, acquisition, and tracking. A satellite still needs enough onboard storage to hold data until it can see a usable receiver. Ground stations need backhaul capacity, data ingest systems, cybersecurity controls, scheduling software, and customer distribution pipelines. A higher-rate link can shift the bottleneck from spacecraft downlink to ground processing, cataloging, and delivery.

EO operators also face a demand mismatch. Some users want full-resolution archives for scientific analysis. Other users want near-real-time alerts, change maps, vessel detections, flood boundaries, fire detections, or site monitoring updates. A high-speed optical link helps both groups, but it helps them differently. For archive users, the link preserves more raw data. For alert users, the link may still be slower than onboard detection if the event must trigger a response before the next downlink opportunity.

The difference can be summarized by comparing the technical function of each approach.

The table shows why optical links reduce the pressure for aggressive onboard processing but do not erase the need for onboard decisions. If the full image is valuable and the link is available, downlinking more data makes sense. If much of the data has little immediate value, or if response time matters more than archive completeness, onboard processing keeps its place.

Cloud Screening and Triage Make On-Orbit Processing Useful

Clouds create a direct case for onboard triage in optical EO. Many optical images contain cloud cover, haze, smoke, or shadow that reduces their immediate value for land, infrastructure, agriculture, or disaster mapping. Ground systems can detect and mask those conditions after downlink, but that means the satellite used storage, downlink time, and ground ingest resources on data that may not help the user. ESA’s Φsat mission addressed exactly that issue by filtering less useful images so that only more usable data came back to Earth.

Cloud screening does not require a satellite to create a finished EO product in orbit. It can act as a first filter. The satellite can classify a scene as likely useful, partly useful, or low priority. It can send quick-look thumbnails, metadata, or quality flags first, then transmit the full scene later if capacity allows. That is less risky than discarding data completely, and it fits missions that need to balance archive value with delivery speed. Triage can be conservative, meaning the system rejects only data that is plainly unusable.

The same logic extends to empty-scene filtering. A maritime surveillance mission may image large ocean areas that contain no vessels. A disaster response mission may cover territory outside the active flood boundary. A border or infrastructure monitoring mission may revisit areas where nothing changed. If onboard processing can detect that a scene has no relevant activity, it can lower the priority for downlink. That does not require the satellite to make a legal, scientific, or operational decision. It only needs to improve the queue.

SAR complicates the picture because it works through darkness and many weather conditions, and it often produces large data files. SAR operators may still benefit from onboard preprocessing or event detection because raw radar data requires specialized handling. Commercial SAR firms such as Capella Space emphasize rapid, taskable, all-weather intelligence, and that market framing points toward faster delivery and faster interpretation, not only better collection.

Onboard processing gains value when the satellite can reduce false urgency. If every image receives the same downlink priority, operators need more communications capacity and larger ground processing pipelines. If the satellite can identify a wildfire plume, a ship-like target, a new flood extent, or a change in an industrial site, it can send a small alert package before the full dataset. That alert might include coordinates, confidence values, acquisition time, viewing geometry, and a small image chip.

A conservative onboard system can also protect science value. It does not need to delete raw data. It can store full data for later transmission, send priority subsets early, and mark remaining data for standard delivery. This pattern makes orbital computing a scheduling and triage tool, rather than a replacement for ground processing.

Latency Creates Different Choices for Civil, Commercial, and Defense Users

Latency has different value depending on the user. A climate scientist may need consistent, calibrated datasets over years or decades. A disaster management agency may need flood, fire, or storm information within hours. A commercial site-monitoring customer may care about same-day delivery. A defense and security user may prefer a detection message within minutes, even if the full image follows later. NASA’s Earthdata latency guidance treats latency as a mission and product property, not a single technical number.

High-speed optical links are strong for bulk latency reduction. They let operators send larger scenes, wider spectral datasets, higher revisit imagery, and less-compressed products. Planet’s SkySat documentation describes a high-resolution constellation of 15 satellites with frequent revisit capability and 50-centimeter orthorectified imagery. That kind of commercial service gains from faster downlinks because more high-resolution data can reach ground workflows sooner.

Orbital computing is stronger for decision latency. A satellite that detects a fire front, oil slick, moving vessel, or runway change can transmit an alert through a lower-bandwidth link before it sends the full image. The alert may carry less information than a complete product, but it can reach a response system earlier. That distinction matters in emergency management and defense and security operations because the first notification may trigger another satellite collection, aircraft tasking, field deployment, or analyst review.

The NASA-ISRO Synthetic Aperture Radar mission shows the value of ground-side science products in a different latency class. NASA describes NISAR as a mission designed to monitor land, water, vegetation, and ice, with global observations of land and ice-covered regions every 12 days from both ascending and descending orbits. NISAR launched on 30 July 2025 from the Satish Dhawan Space Centre in India, and NASA’s May 2026 mission updates described active data applications work and new observations, including mapping related to Mexico City subsidence.

Civil government programs often need trust, continuity, and access more than instant detection. The Copernicus Data Space Ecosystem provides open access to Copernicus Sentinel mission data and related services, reflecting a shift toward scalable access and processing for large user communities. In that setting, optical downlinks, cloud processing, archives, and user platforms all matter. Onboard processing can help, but the public value often depends on transparent, repeatable ground products.

Defense and security users face a different balance. They may accept a machine-generated alert as an early cue, then ask for full-resolution data after the fact. They may also prefer onboard triage when communications links are contested, ground stations are unavailable, or data must move through restricted channels. Optical communications can offer high data rates and narrow beams, but they still require terminals, weather access for ground links, and network planning. Neither architecture removes the need for mission-specific security engineering.

Ground Processing Still Wins for Calibration, Archives, and Scientific Traceability

Ground processing remains the stronger choice for science-grade EO products because calibration, validation, correction, reprocessing, and archive management benefit from larger computers, stable software, expert oversight, and access to reference datasets. NASA’s data processing levels explain how data products move from raw instrument data to higher-level products that users can interpret. Moving from Level 0 to Level 4 often requires atmospheric correction, geolocation, quality assessment, compositing, and modeling that ground systems can manage more transparently.

The Landsat Collection 2 program shows why archives matter. The U.S. Geological Survey describes Collection 2 as a major reprocessing effort that includes Level 1 data from Landsats 1 through 9 and Level 2 and Level 3 science products from Landsats 4 through 9. Reprocessing improves consistency across decades. A spacecraft computer cannot replace that archive-level work because the value comes from uniform calibration across missions, sensors, software versions, and time.

Onboard processing also raises auditability concerns. If a satellite discards data before downlink, later users may never know what was lost. If it transforms data into a reduced product, users may need detailed information about algorithms, model versions, thresholds, compression settings, and false-positive behavior. Science users need confidence that a detected trend comes from Earth, not from a changing onboard model. Commercial and legal users may need evidence chains that explain how a product was made.

Optical downlinks support traceability by preserving more raw or lightly processed data. If a mission can downlink full instrument data at acceptable cost and latency, it can keep more options open. Ground users can later apply improved algorithms, combine data with other sensors, reprocess historical archives, and correct early mistakes. That matters for climate records, land-change analysis, hydrology, agriculture, and infrastructure monitoring.

Onboard computing still has a place in science missions, but its best role may be modest. It can compress data, detect instrument problems, prioritize rare events, or create browse products. It can run conservative quality checks, not final science interpretation. For missions that need open data, long time series, and public reproducibility, onboard processing should avoid irreversible filtering unless the mission has strong validation and clear user consent.

Commercial EO has more freedom. A private operator selling alerts rather than raw imagery may decide that onboard analysis is the product. That can work when customers buy speed and relevance instead of full scientific traceability. Even in that case, the operator benefits from retaining enough raw data to troubleshoot, retrain models, prove performance, and serve customers who later need full evidence.

Hybrid Architectures Fit Most Earth Observation Business Models

A hybrid architecture uses high-speed links for bulk data movement and onboard computing for time-sensitive triage. This model fits most EO business cases better than a pure choice between space computing and optical downlink. It recognizes that satellites gather data under changing conditions, customers assign different value to speed and completeness, and ground networks have uneven capacity. ESA’s EDRS and NASA’s optical demonstrations show how communications capacity can grow, and ESA’s Φsat missions show that onboard artificial intelligence can remove low-value data or generate early insight.

Hybrid design starts with a hierarchy. Raw data gets preserved when possible. Quick-look products move early. Alerts move first when the satellite detects an event. Full-resolution data follows through optical downlink, radio frequency downlink, relay satellite, or later ground pass. Ground systems complete calibration, distribution, and archive work. This division gives each layer a job that fits its strengths.

A commercial constellation might use onboard computing to rank image chips by customer interest. For example, a ship-detection service can downlink a compact alert and a small radar chip first, then downlink the full SAR scene through a high-capacity link. A crop-monitoring service can avoid giving premium downlink priority to clouded optical scenes. A disaster-monitoring service can send a first assessment early, then replace it with a ground-processed product after calibration and analyst review.

Hybrid architecture also fits cost control. Radiation-tolerant computing hardware costs more than ordinary ground servers and usually offers less computing performance. Space-qualified processors, storage, thermal control, power budgets, software assurance, and fault tolerance all add cost. Ground computing can scale through cloud platforms and data centers, and it can be updated more easily. The satellite should process data only when the value of earlier filtering or earlier detection exceeds the cost and risk of putting that function in orbit.

Optical communications also have cost and coverage limits. A mission needs optical terminals, ground stations, scheduling, atmospheric diversity, and weather-aware routing. Relay systems reduce some direct-to-ground limits but add service cost and dependency on another network. For many missions, the rational design is not maximum onboard computing or maximum downlink capacity. The better design assigns routine bulk transfer to high-speed links and assigns immediate relevance detection to onboard software.

EO companies can use this architecture to sell different service tiers. Archive users can pay for full datasets. Alert users can pay for low-latency notifications. Government users can require retention, audit logs, and transparent processing. Defense and security users can request restricted delivery paths and early cueing. A single satellite architecture can support these tiers if it treats onboard compute as a prioritization layer rather than a universal replacement for ground processing.

The Investment Question Is Where Value Gets Captured

Investment in orbital computing makes sense when the satellite creates value before the data reaches Earth. That value can come from faster alerts, lower downlink cost, reduced storage needs, improved tasking, or better use of scarce ground contact time. Investment in optical downlinks makes sense when preserving and moving large data volumes creates more value than filtering them. The business decision turns on customer behavior, not only engineering performance.

Commercial EO firms increasingly sell answers, not imagery alone. Customers may want vessel locations, damage assessments, crop health indicators, construction progress, ground movement, or change alerts. If the product is an answer, onboard computing can help when the answer is time-sensitive and simple enough for reliable space-based inference. If the product is evidence, analysis, or an archive, high-speed downlink and ground processing may matter more.

The economics also depend on revisit rate. A satellite that images a site once every several days may have less need for onboard triage because the data volume is manageable. A constellation that revisits high-demand areas many times a day may benefit more from onboard filtering. Capella Space advertises revisit times under three hours over selected regions, and Planet SkySat supports frequent high-resolution tasking. Dense revisit strategies make prioritization more valuable because the system collects more candidate observations than customers may need.

Insurance and liability matter too. If a satellite’s onboard model misses an event, the operator may need to explain why. If the model sends false alerts, the customer may lose trust. This makes onboard processing most attractive where outputs can be framed as cues, rankings, or preliminary alerts. Final claims, legal evidence, scientific products, and high-cost operational decisions often need ground-side review, validation, or human approval.

Regulation can shape the choice. Remote sensing rules, export controls, spectrum licensing, cybersecurity requirements, defense contracts, and data-sovereignty rules can influence where data gets processed and who can access it. Onboard processing might reduce the movement of sensitive raw data, but it can also create new compliance questions if algorithms classify objects or produce intelligence products in orbit. Optical links can move more data, but they require ground sites, network routing, and secure handling at scale.

Financing choices follow these risks. Optical terminals and ground networks are infrastructure investments. They can serve many satellites and many customers if standardized. Onboard computing is more mission-specific. It can create product differentiation, but it may become obsolete faster as artificial intelligence models and customer demands change. Operators need upgrade paths, software assurance, and enough raw data retention to retrain or audit their systems.

Practical Design Rules for Operators and Users

Operators should begin with the product promise. If the promise is a trusted data archive, the spacecraft should preserve data quality and use high-capacity downlinks where practical. If the promise is low-latency alerts, the spacecraft should carry enough onboard processing to detect and prioritize events. If the promise combines both, the architecture should separate alert generation from final product creation.

A useful rule is to process in orbit only when delay changes the value of the data. Cloud filtering, empty-scene detection, event cueing, and priority scoring meet that test. Full atmospheric correction, long-term trend analysis, multi-sensor fusion, scientific reprocessing, and legal-grade evidence packages usually do not. They need ground systems because they draw on reference data, external models, analyst review, or archive-wide consistency.

Operators should avoid irreversible deletion unless the business case is clear and users accept the loss. A satellite can mark low-quality data as low priority rather than discard it. It can send small preview products first and preserve the full data for later transmission. It can use adaptive compression based on scene content. These choices reduce risk because they keep ground users from losing potentially useful data.

Users should ask what they are buying. A raw image, a calibrated image, a map layer, a detection, and an alert are different products. A buyer that needs court-admissible evidence, engineering-grade measurement, or scientific reproducibility should favor traceable ground processing and raw data retention. A buyer that needs a fast cue for an analyst or response team may accept onboard detection as an early warning layer.

Procurement language should specify latency at the product level. The delivery time for a raw file, a preview image, a machine-generated alert, and a validated product may differ by minutes or hours. A contract that says “low latency” without defining the product can create confusion. Better specifications state acquisition-to-alert time, acquisition-to-full-product time, confidence thresholds, false-alarm handling, archive retention, and reprocessing rights.

Standards and interoperability also matter. EO users increasingly depend on catalogs, application programming interfaces, cloud-native data formats, and processing platforms. Faster downlinks will have limited value if data cannot be found, accessed, and used quickly after arrival. Onboard computing will have limited value if its outputs do not fit user workflows. The satellite is only one part of the data service.

High-Speed Optical Links Do Not Eliminate On-Orbit Processing

High-speed optical links reduce one of the best-known arguments for on-orbit processing: limited downlink capacity. NASA’s TBIRD and LCRD-related demonstrations, ESA’s EDRS, and commercial optical communications work all point toward a future in which many EO missions can move more data faster than older radio frequency architectures allowed. That does not mean every byte should automatically go to the ground first. The value of early decisions remains when contact windows, weather, customer urgency, or mission security constrain delivery.

On-orbit processing survives because the EO market is moving from collection toward service delivery. A satellite that can say “nothing changed,” “clouds block this scene,” “a vessel-like object is present,” or “this area deserves priority downlink” can improve the whole data chain. The computation does not need to replace ground analytics. It only needs to make communications, storage, and attention more efficient.

High-speed links also make onboard computing better. If a satellite can transmit both a quick alert and the full scene, users get speed and evidence. If a ground system can send updated models to the satellite through a capable link, onboard software can improve over time. If relay networks reduce gaps between acquisition and downlink, onboard processing can shift from heavy filtering to smarter prioritization.

The strongest architecture for many EO missions as of 15 May 2026 is a layered one. The satellite collects data, performs conservative triage, and sends early alerts or previews when valuable. Optical links move full-resolution data as soon as conditions allow. Ground systems produce calibrated, traceable, and archive-ready products. Customers receive the level of speed, evidence, and cost that fits their decision.

Orbital computing for Earth observation satellites should not be judged as a rival to high-speed optical downlinks. It should be judged as a way to spend communications capacity more intelligently. Optical links move more data. On-orbit computing decides what deserves attention first. The best EO systems will use both, with the split driven by latency, trust, cost, and the customer’s tolerance for early machine judgment.

Summary

The debate over orbital computing and optical downlinks often starts from the wrong assumption. It treats the issue as a contest between processing data in space and sending data to Earth. EO missions do not operate that neatly. They collect data under physical constraints, transmit it through scheduled networks, process it through complex ground systems, and serve users who value speed, completeness, confidence, and cost in different proportions.

High-speed optical links weaken the case for aggressive onboard filtering because they make it easier to preserve full-resolution data. That is particularly important for science missions, public archives, calibration programs, climate records, and commercial customers that need evidence-grade products. When data can be downlinked quickly and affordably, ground processing remains the better place for most heavy analysis.

On-orbit processing remains valuable where time or capacity changes the product. Cloud screening, event detection, compression, target triage, and priority scoring can improve the EO chain before the full dataset reaches Earth. The strongest use cases treat onboard outputs as early cues or routing decisions, not as final products that replace ground validation.

The practical answer is hybrid architecture. Satellites should do enough processing to reduce waste, cue urgent events, and rank data for downlink. Optical links should move as much full-quality data as possible. Ground systems should handle calibration, archives, product generation, and deeper analysis. That split preserves scientific trust, supports commercial speed, and gives defense and security users faster awareness without forcing every mission into the same technical model.

Appendix: Useful Books Available on Amazon

• Remote Sensing and Image Interpretation
• Introduction to Remote Sensing
• Remote Sensing of the Environment
• Remote Sensing: Models and Methods for Image Processing
• Satellite Communications Systems

Appendix: Top Questions Answered in This Article

Does High-Speed Optical Downlink Remove the Need for On-Orbit Processing?

No. High-speed optical downlink reduces the need to process data in orbit only for missions that can transmit full datasets quickly, affordably, and reliably. On-orbit processing still helps when data needs triage, alerts, compression, cloud screening, or priority routing before a complete ground product is available.

Why Would an Earth Observation Satellite Process Data Before Downloading It?

A satellite may process data before download to reduce wasted bandwidth, identify high-priority scenes, detect time-sensitive events, or send compact alerts ahead of full imagery. This is most useful when users value fast awareness more than immediate access to a complete calibrated product.

Why Are Optical Links So Important for Earth Observation?

Optical links can move much larger data volumes than many radio frequency systems under suitable conditions. That matters because modern Earth observation satellites can collect high-resolution, multispectral, hyperspectral, radar, or video data that strains older downlink architectures.

What Is the Main Weakness of Optical Downlinks?

Optical downlinks need accurate pointing and suitable atmospheric conditions for direct-to-Earth service. Clouds and weather can interfere with optical ground reception, so operators often need geographically distributed ground stations, relay options, storage, and fallback communications plans.

What Types of Earth Observation Products Still Need Ground Processing?

Science-grade products, climate records, calibrated imagery, and archive products usually need ground processing. These products depend on validation, repeatable algorithms, reference datasets, quality control, and long-term consistency across sensors and missions.

Where Does Artificial Intelligence Fit in On-Orbit Processing?

Artificial intelligence can help satellites classify scenes, detect clouds, find objects, identify changes, or rank data for downlink. Its best early use is conservative triage, where the system improves data handling without making irreversible decisions that users cannot review.

Is On-Orbit Processing Better for Defense and Security Uses?

It can be better for early cueing and restricted workflows because some defense and security users need faster indications of activity. Full-resolution data, analyst review, secure ground processing, and audit trails often remain necessary for higher-confidence decisions.

Could Satellites Downlink Everything and Process It Later?

Some missions may be able to downlink nearly everything if they have enough optical capacity, ground access, storage, and budget. Many missions still face scheduling, weather, customer-latency, and data-management limits that make selective onboard triage useful.

What Is the Best Architecture for Commercial Earth Observation?

The strongest commercial architecture often combines onboard triage, high-speed optical downlink, and ground analytics. This allows the operator to send alerts quickly, preserve full evidence when needed, and create higher-value products through ground processing.

What Should Buyers Ask Earth Observation Providers?

Buyers should ask whether delivery promises refer to raw data, preview images, alerts, or validated products. They should also ask about latency, confidence thresholds, false alarms, raw data retention, reprocessing rights, data security, and whether onboard processing can be audited.

Appendix: Glossary of Key Terms

Orbital Computing

Orbital computing means processing data aboard a spacecraft rather than sending all raw data to Earth first. In Earth observation, it often supports compression, quality screening, change detection, event alerts, or tasking decisions before full ground processing occurs.

Earth Observation

Earth observation refers to measuring Earth’s land, oceans, atmosphere, ice, and human activity from satellites, aircraft, drones, ships, or ground sensors. Satellite Earth observation often uses optical, infrared, radar, or microwave instruments to create images and measurements.

Data Latency

Data latency is the time between a sensor acquiring data and a user being able to access a usable product. In Earth observation, latency includes collection, storage, downlink, ingest, processing, quality checks, cataloging, and distribution.

Optical Communications

Optical communications use light, often near-infrared laser beams, to transmit data. Space optical links can offer much higher data rates than many radio systems, but they require accurate pointing and suitable atmospheric or relay conditions.

Radio Frequency Communications

Radio frequency communications use electromagnetic waves at radio frequencies to send data between spacecraft and ground stations. They have long heritage in spaceflight and can work under conditions that may block optical links, but bandwidth can be more constrained.

Synthetic Aperture Radar

Synthetic aperture radar is an active radar imaging technique that sends microwave pulses toward Earth and measures the returned energy. It can operate at night and through many weather conditions, making it valuable for maritime monitoring, land deformation, flooding, and security applications.

Cloud Screening

Cloud screening is the process of identifying clouds, haze, shadow, or other atmospheric interference in optical imagery. Onboard cloud screening can lower the priority of imagery that has limited immediate value for land, agriculture, or infrastructure monitoring.

Triage

Triage means ranking or sorting data by value, urgency, or quality. In satellite operations, triage can decide which scenes, image chips, alerts, or metadata should be transmitted first during limited contact opportunities.

Data Relay System

A data relay system passes information from one satellite through another satellite before it reaches the ground. This can reduce dependence on direct contact between a low Earth orbit satellite and a ground station.

Ground Processing

Ground processing refers to the software, computing, validation, and distribution work performed after satellite data reaches Earth. It can convert raw observations into calibrated images, measurements, map layers, alerts, archive products, or analysis-ready datasets.