Inside Swiss Re's Drone Flood Patent: A Vulnerability Curve for Every Building

A drone flies a user-drawn polygon, photographs every point inside it from at least two positions, and the system triangulates a bare-earth terrain model with a ground sampling distance of 5 cm or less, with ground control points holding root mean square error to 15 to 40 mm. That measurement chain is the independent claim of US Patent 12,705,882 B2, issued to Swiss Reinsurance Company Ltd on August 11, 2026 (USPTO). The actuarial payload sits at the far end of it.

Key Takeaways

• Claim 18 asks for separate vulnerability curves, one per building and one per sub-unit inside the polygon, each indexed to inundation height or return period. Flood ratemaking has run for fifty years on zone assignment.
• Claim 7 supplies a parametric hook, matching structures against a warehouse of historically measured flood damage at corresponding inundation heights. That is a payout basis, and it relocates residual basis risk from zone assignment to depth measurement.
• The single independent claim is a system claim, reciting drones, sensors, a calibration module and a filter rather than a bare method, which is a materially harder abstract-idea rejection to write after Recentive Analytics.
• $3.4 billion of global insured flood losses in 2025 against $107 billion of total insured natural catastrophe losses (Swiss Re Institute), with an estimated 83% flood protection gap across Asia.

Patent Details

From Overlapping Photographs to a Bare-Earth Model

The operator draws a polygon on a topographic map through a geo-tagging interface. Drones image the enclosed area so every possible location is captured in at least two photographs from different positions, with image overlap the specification puts at 70 percent or more. Common points are identified across image pairs, a ray is generated from the camera to each measured ground point, and the intersection of those rays fixes a three-dimensional coordinate.

Georeferencing comes from ground control points, physical marks tied to surveyed coordinates. The claim sets the tolerance directly: "to achieve an accuracy of 5 cm or less for elevation and surface inclination drone-based measurements, root mean square error values at the at least two ground control points is in a range of 15-40 mm." Dependent claims add post-processing and real-time kinematic GPS correction, and claim 14 escalates the number of control points until the tolerance is met.

The last step is a progressive morphological filter, tuned by parameter calibration, that strips vegetation and structures out of the surface model to leave a digital terrain model. That distinction is not cosmetic. Water routes over bare earth, and a canopy-contaminated surface model will put flow paths in the wrong place.

Swiss Re filed the priority application in Switzerland in October 2022, carried it through PCT a year later, and entered the United States in May 2024. That the single independent claim is drafted as a system rather than a method is doing work: insurance analytics claims have been the softest target under Section 101 since Alice, and a claim reciting a measured error tolerance at physical ground control points is considerably harder to reject as an abstract idea than one reciting a calculation.

Sub-Building Curves and Where the Basis Risk Goes

The output layer is where the pricing content lives. A flood hazard aggregator generates vulnerability measures for each unit and sub-unit inside the polygon, keyed to forecasted or user-selected return periods and inundation heights, and the specification runs those through to monetary property damage and business interruption estimates. Each sub-unit carries its own damage curve, with a dependency scheme for cascading failures.

That is a different object than a property-level flood score, and different again from the zone-based rating the National Flood Insurance Program spent decades on. A single-building vulnerability curve assumes one depth applies to the whole structure.

A model resolving terrain to 5 cm can distinguish a loading dock two feet below the finished floor from the floor itself, and can price the sub-unit that floods first separately from the one that floods last. For a commercial site where the switchgear sits in a basement and the production line sits on a slab, those are not the same exposure and have never been priced as different ones.

Claim 7 supplies the parametric hook. The aggregator determines damage by matching structures against a digital data warehouse of historically measured flood damage at corresponding inundation heights and return periods, which is a payout basis rather than an indemnity assessment. It relocates residual basis risk from zone assignment to depth measurement, and that is a different and smaller error term.

The distinction matters for how a pricing actuary should treat the improvement. Zone-based basis risk is categorical: a property is inside or outside a mapped zone, and everything inside shares a rate. Depth-based basis risk is continuous and, on this measurement chain, bounded by a stated tolerance. Trading a categorical error for a bounded continuous one narrows the distribution of pricing error without eliminating it, and it does so at the sub-unit level where the loss actually occurs.

What the Measurement Chain Cannot Reach

The precision is real and it stops at the terrain. Everything above the bare-earth model, the return period, the flood frequency distribution, the correlation between one site and the next, is still modeled rather than measured, and none of it improves because the elevation data got sharper. A vulnerability curve indexed to inundation height is only as good as the estimate of how often that height occurs, which comes from hydrology and climate assumptions the drone never touches.

That is the asymmetry a reinsurer buying this capability inherits. The measured half of the calculation gets a stated tolerance of 15 to 40 mm. The modeled half carries the same parameter uncertainty it always did, and a precise damage curve applied to an uncertain frequency produces a precise-looking number whose error is dominated by the half nobody measured. Presenting a 5 cm terrain model alongside a return-period assumption invites more confidence in the combined output than the weaker input supports.

The economics point the same direction. Flying a polygon, placing surveyed ground control points and escalating them until tolerance is met is a per-site exercise, not a portfolio one. It is affordable against a large commercial risk, an industrial campus, a data centre, a port terminal, where a single site carries enough premium to justify a survey. It does not scale to the residential book where the exposure count is highest.

That is where the protection gap sits. Global insured flood losses ran $3.4 billion in 2025 against $107 billion of total insured natural catastrophe losses (Swiss Re Institute), and the gap across Asia runs near 83%. The residential side is where independent modelling has found the most exposure outside mapped zones, work First Street Foundation built its national flood model to quantify.

The places with the widest gap are the places least able to fund a per-site survey and least likely to have surveyed ground control points already in the ground. Private flood carriers have been growing share against that gap for several years, but on portfolio models rather than per-site measurement. A technology that prices the best-documented risks more accurately narrows uncertainty where the market already functions, and leaves it untouched where the market does not.