Enter, exit, dwell
Evaluate moving entities against areas of interest, require dwell duration, suppress duplicate alerts and retain the triggering observation.
Operational geospatial services
Observe. Decide. Act in space and time.
The engine joins live feature streams, event rules, video telemetry, mission workflows and 3D scene services in one data plane instead of isolating them behind separate server products.
Real-time data plane
Normalize first. Route and analyze second.
Ingest adapters translate heterogeneous sources into common feature and event representations. Consumers can subscribe to current state or updates without understanding every upstream protocol.
InputsTAK / Cursor-on-TargetOGC SensorThings / SOS / SensorMLMQTT · Redis · Kafka RESTHTTP polling · cron SQL · WebSocketGNSS · AIS · ADS-B · fleet feeds
→
State + eventsStream feature storeCross-subsystem event busGeofence and threshold rulesDwell, velocity and pattern detectorsSQL-driven DuckDB data triggersWatchlists, alerts and audit records
→
OutputsEsri-compatible StreamServerWebSocket operational updatesFeatureServer / OGC feature viewsDashboards and alertsCoT rebroadcast and webhooks
Event processing
Rules that understand geometry and movement.
Event handling spans simple thresholds and operational patterns, with state retained for alerts, review and after-action analysis.
Value crossing
Detect sensor values crossing configured bounds, apply cooldowns and route notifications through SMS, email, push or HMAC-signed webhooks.
Movement events
Classify starts, stops, turns, U-turns, acceleration and deceleration; calculate dwell, rendezvous and co-travel patterns.
Space-time analytics
Aggregate H3 space-time cubes, compare periods, score emerging hotspots and build time-series bins from multi-source observations.
Watchdog and degradation
Track subsystem health and escalate through normal, degraded, critical and emergency tiers when dependent services fail.
Cross-feed matching
Match entities by equality, containment, regular expression or geographic proximity with confidence scoring across feeds.
Reactive SQL over attached files
Author SQL predicates against watched DuckDB data files; the events extension re-evaluates rules on each commit and fires alerts through the same notifier—with a poll fallback when the extension is unavailable.
Video service stack
Video is geospatial data.
Streams and files are joined to telemetry, map position, events and analytical results rather than living in an isolated player.
ING
Multi-protocol ingest
RTSP, RTMP, HLS, DASH and file-based video workflows, with HTTP range support for seeking and operational playback.
KLV
Full-motion video metadata
Extract MISB 0601 KLV telemetry, GoPro GPMF metadata and DJI SRT GPS to position platforms, frames and sensor footprints.
GEO
Pixel-to-ground mensuration
Use camera models and DEMs to project image coordinates onto the ground and persist video digitizing as map features.
AI
Detection and tracking
YOLO-based object detection, OCSort tracking, zone entry/exit and dwell analysis, plus enhancement and frame extraction.
AAR
Replay and evidence
Synchronize tracks, map events and video observations for after-action review instead of treating each feed as a separate timeline.
3D service engine
Serve scenes in the formats clients already understand.
3DGeoServEngine handles both Esri I3S/SceneServer and OGC 3D Tiles pipelines, alongside models, point clouds, BIM and emerging volumetric formats.
| Data family | Ingest / processing | Serving / client path | Analytical use |
|---|---|---|---|
| I3S / SLPK | Read scene layers and SLPK packages; expose nodes, geometry, textures and statistics | Esri-compatible SceneServer and ArcGIS SceneLayer; bridge to Cesium-oriented visualization | Feature queries, styling and scene inspection |
| OGC 3D Tiles | Read 3TZ/tilesets; generate optimized 3D Tiles 1.1 and implicit tiling | Cesium-compatible tileset and binary-content endpoints | Tile metadata, style evaluation and feature selection |
| LiDAR / point cloud | LAS, LAZ, E57 and COPC-oriented conversion/streaming paths | Range-addressable point-cloud content and 3D Tiles outputs | Elevation, classification, terrain and surface workflows |
| Gaussian splats | PLY, SPLAT and SPZ ingestion with compression/conversion | 3D Tiles 1.1 or raw range-served splat content | Rapid reality capture and immersive inspection |
| BIM / urban models | IFC2x3/IFC4, CityGML, CityJSON, GLB/glTF and model conversion | Web 3D viewers, scene overlays and tile conversion | Indoor hierarchy, routing, assets and digital twins |
| Terrain / elevation | COG, DTED, SRTM and generated DEM products | ImageServer/LERC elevation, raster DEM terrain and 3D analysis services | Viewshed, slope, skyline, volume and line-of-sight |
See it in motion
Real product capture
Full-resolution capture of the client actually running.
The same 3D terrain and measurement tools shown above, recorded live against the real client rather than staged for a demo.
Mission and application services
Operational workflows beyond visualization.
The shared backend includes mission planning, field data collection, incident command and application-composition capabilities that would normally span multiple server roles.
Versioned mission packages
Role-protected packages, immutable approved baselines, map views, operational overlays, comments, actions, audit trails and DOCX/PDF/mapbook publication.
Offline incident command
Incident records, FEMA/NIMS ICS forms, IAPs, resources, tasks, assignments, geofences, documents, dashboards and government-standard CAP/EDXL/NIEM/CoT composition.
Forms and sync
XLSForm/Survey123-style forms, Field Maps formInfo, repeat groups, attachments, GlobalIDs, offline dirty tracking, conflict handling and FeatureServer/PostGIS synchronization.
Experience and dashboard builder
Compose responsive multi-page applications from map, filter, table, chart, indicator and navigation widgets, wire interactions and export a standalone HTML application.
Standard grids for shared reference
MGRS, USNG, WGS84 and GARS overlays.
Cross-agency and military operations need a shared grid vocabulary, not just a basemap. The four grid systems below render over the live map alongside the operational layers documented on this page.
