Coverage Intelligence (Beta)
Know what your drones actually saw, not just where they flew.
The Problem
Flight logs tell you where a drone went. They don't tell you what its camera observed. A drone can fly directly over a gate checkpoint while its camera points in the opposite direction — and today, there's no way to know.
Security operators with SLA commitments face this gap every day. When a customer asks "was my perimeter inspected last night?", the best answer available is flight count and total hours — a proxy metric that says nothing about actual visual coverage. There's no spatial proof, no way to identify blind spots, and no evidence that critical checkpoints received the attention they were promised.
Coverage Intelligence closes that gap.
Before You Begin
Coverage Intelligence computes the camera's ground-projected field of view from flight log telemetry and visualizes it on a map. Instead of a flight path line, you see the actual area the camera observed — as a coverage overlay showing what was seen and what was missed, or as a dwell time heatmap showing how long each area received camera attention.
You define what matters — your site boundary, perimeter fences, critical checkpoints — and the system continuously evaluates whether those areas are getting the coverage your SLA requires. When it's time to report, you generate a PDF that shows your customer exactly what was inspected, how thoroughly, and where the gaps are.
How to Access
Navigate to your organization URL with /coverage-intelligence/ appended: https://.flytbase.com/coverage-intelligence/
The trailing slash is required. Without it, the page won't load. A dedicated navigation entry is coming — for now, use the direct URL.
Setting Up Your Site
An in-product setup wizard runs on first visit to a new site and walks you through four steps.
Draw your site boundary. This is the area you're contractually responsible for — the polygon against which overall coverage percentage is measured. Set a target (e.g., 90% coverage) and the system tells you whether you're meeting it.
Mark your critical points. Drop POI pins at locations that require specific attention — gates, parking lots, generator rooms, high-value assets. Each gets a name, a dwell time target (how long the camera should observe it), and optional inspection notes describing what to look for. These are the checkpoints your SLA probably names explicitly.
Trace your fence lines. Draw polylines along perimeter fences that need regular inspection. The system measures what percentage of each fence was visually covered and highlights exactly where the blind spots are — down to the meter.
Confirm your rolling windows. The system auto-creates Daily (24h), Weekly (7d), and Monthly (30d) evaluation windows. These define the time periods your coverage is measured against. Customize them to match your reporting cadence — add a 12-hour shift window, a bi-weekly cycle, whatever fits your operation.
Once setup completes, if historical flight data exists, the system immediately runs a coverage analysis so you see real results on day one.
Coverage Visualization
Select a site, choose your drones and time range, and hit Compute All. The system processes every flight in the window and renders the results on the map.
Frequency mode renders a dwell time heatmap. Cool colors indicate areas the camera passed over briefly (2–4 seconds in transit). Warm colors indicate sustained inspection attention (20–60+ seconds of observation). This naturally separates transit corridors from actual inspection work — no filtering needed.
Hover over any grid cell to see its accumulated dwell time, how many flights contributed coverage, and when it was last observed. Click a cell to trace back to the specific flights that covered it.
The grid resolution is adjustable via a slider (5m to 50m cells, default 10m). Smaller cells give precision; larger cells give a cleaner overview and better performance over long time ranges.
Compliance Tracking
With your annotations defined and coverage computed, the system evaluates compliance across all your rolling windows simultaneously.
Site boundary: Percentage of your boundary area that received camera coverage within the window. Compared against your target with a clear met/at-risk/missed indicator.
Fences: Linear coverage percentage for each fence — what fraction of the fence line was observed. Blind spots (consecutive sections with zero coverage) are identified by location and length. In frequency mode, the fence line itself is colored by dwell time — red sections got sustained attention, gaps got none.
Points of interest: Accumulated dwell time for each POI compared against its target. A POI with a 60-second daily target that only received 12 seconds of observation is flagged immediately.
Reports
Coverage data is only useful if you can share it with your customer. The reporting system generates PDF documents from your computed coverage.
Create report templates that define which rolling window to evaluate, which annotation types to include (boundary, fences, POIs, incidents), and who receives them. A "Daily Client Report" might cover boundary and fence compliance for the last 24 hours. A "Weekly Internal Review" might include everything.
Important: The report captures whatever coverage data is currently computed on the map. Compute your coverage for the relevant time range first, then generate the report. If nothing is computed, the report will show annotations but no heatmap overlay.
Each report includes a coverage map snapshot with heatmap overlay, per-annotation compliance tables with met/missed indicators, a flight summary (total flights, hours, pilots, drones), and any incidents logged during the period. The format is designed to go directly to an enterprise security customer as proof of patrol coverage.
AI Agent
An AI assistant is embedded in the dashboard. It can answer questions about your coverage data, check compliance status, and help you navigate the platform.
Ask it things like "Which POIs missed their target today?", "Show me all areas with zero coverage this week", or "Help me set up reports." The agent reads your live dashboard data and can manipulate the map — zooming to problem areas, toggling layers, switching views — to show you answers visually, not just in text.
The agent is also aware of the platform's features and can guide you through workflows you haven't tried yet.
Current Limitations
Coverage Intelligence is in beta. A few things to know:
Gimbal data is approximated. The flight log API does not yet include gimbal pitch and yaw telemetry. The system assumes a fixed 45° downward camera angle with the gimbal facing the same direction as the drone. This is a reasonable approximation for typical patrol operations, but means the coverage footprint is estimated, not exact. When gimbal telemetry is added to the flight log pipeline, the system will use actual camera orientation — no action needed on your end.
Wide camera only. Thermal and zoom camera FOVs are not computed in this version. Coverage represents what the wide-angle camera observed.
Flat terrain assumed. Altitude is measured relative to the docking station. On sites with significant elevation changes, coverage projections may be slightly over- or under-estimated in hilly areas.
What's Next
Real gimbal telemetry is the highest-priority improvement — it will make coverage projections significantly more accurate. Thermal and zoom camera support, scheduled automatic report delivery, and cross-site analytics are on the roadmap.
To Get Started
You can visit the following tutorial videos:
Last updated