LaunchDetect is geospatial intelligence for space launches and orbital activity. Real-time thermal infrared detection from three geostationary satellites, fused with the satellite catalog, NOTAMs, AIS, ADS-B, and wildfire data — all geocoded and queryable.
NOAA GOES-18, GOES-19, JMA Himawari-9 ABI band 7 (3.9 µm). Continuous frame cadence. 30-90s ignition-to-verdict.
17 active spaceports with pad-level coordinates. Cross-referenced against every thermal anomaly.
CelesTrak TLE refresh. 10,000+ objects in LEO/MEO/GEO/HEO. Categorical filters and pass prediction.
FAA AST notices to air missions. Range-safety exclusion polygons, downrange recovery corridors.
Real-time commercial and general-aviation traffic. Cross-referenced with NOTAM exclusion zones.
SpaceX droneship tracking (JRTI, OCISLY, ASOG), sea-based launch platforms, range-safety vessels.
Same sensor stack as launch detection, repurposed for fire detection. Public dashboard at /fire.
JSON, RSS, per-launch citation pages at stable URLs. Public data-library archive.
LaunchDetect's space GIS is geospatial-first. Thermal anomalies are extracted on GPU-accelerated pipelines from raw NOAA Level 1b radiance products with ~0.1 K brightness-temperature precision. Each anomaly is geocoded against a curated spaceport registry; only anomalies within a known pad geofence and above thermal+temporal thresholds advance to verdict. Confidence is set conservatively (current plume threshold 0.70) to keep the false-positive rate below 5%.
The architecture is published in detail at launchdetect.com/why/#methodology. The detection pipeline is AWS-native: S3 ingest, GPU Lambda extraction, DynamoDB state, EventBridge orchestration, and CloudFront-delivered citation pages.
Open the live platform →