The DJI problem

Summary¶

DJI makes excellent drones. That is not the problem. The problem is that every DJI flight routes data through servers outside EU jurisdiction, every DJI payload runs on a closed interface, and every DJI operator is dependent on a Chinese company's continued goodwill for the platform's continued function. For a hobbyist, this is an acceptable trade-off. For a Czech municipality doing flood assessment, a Baltic civil defence group, a NATO- adjacent research institution, or any organisation where data sovereignty is a procurement requirement — it is not acceptable at all.

This article explains the gap that created libdrone and why it is now the EU civilian reference implementation for open-source aerial infrastructure.

Concept¶

Why DJI stopped being a default¶

Until roughly 2022, DJI was the unquestioned default for anyone needing an affordable, capable drone. The hardware remained — and remains — genuinely impressive. What changed was the political and regulatory environment around it.

Several converging pressures made DJI problematic for institutional EU use:

Data routing. DJI's app transmits telemetry, imagery metadata, and in some configurations video frames to DJI servers. The destination of that data is outside EU jurisdiction, outside GDPR enforcement reach, and outside any EU institution's ability to audit. For a commercial photographer, this is a privacy inconvenience. For a civil protection officer mapping a flood in a European city, it is a data sovereignty breach.

US restrictions as a signal. The US Department of Defense added DJI to its list of Chinese military companies in 2022. The US government banned DJI purchases across federal agencies. Whatever the specific legal rationale, the signal was unambiguous: US government and NATO-adjacent institutions do not consider DJI a trusted platform. EU institutions operating in NATO-compatible procurement frameworks read this signal correctly.

Supply chain opacity. DJI's component sourcing is not publicly documented. An institution cannot audit what is inside a DJI drone. It cannot verify the firmware running on the flight controller. It cannot know whether a software update changed the data routing behaviour. This is not a theoretical concern — it is a structural property of closed-source hardware that makes independent verification impossible.

Closed payload interface. DJI's payload SDK is gated behind a developer programme with significant cost and NDA requirements. A sensor manufacturer cannot build a DJI-compatible payload without DJI's permission and without committing to terms that include data sharing with DJI. For any institution that wants to deploy its own sensor — an air quality monitor, a radiation survey instrument, a thermal camera — this is a fundamental constraint.

The alternatives that did not fill the gap¶

When EU institutions began moving away from DJI, they found that the alternatives were worse in different ways.

Commercial enterprise drones (Parrot Anafi, Autel, Skydio) addressed the Chinese supply chain concern but kept the closed-software and closed-payload problems. Paying €2,000–8,000 for a drone that still has a proprietary payload interface is not a meaningful improvement for an institution that wants to deploy its own sensors.

Academic research platforms (PX4-based custom builds, institutional UAV programmes) were genuinely open but started at €10,000–50,000 and required institutional procurement processes, specialised technical staff, and multi- year deployment timelines. A Czech municipal civil protection office with a €2,000 budget and no dedicated drone engineer cannot use these.

Racing FPV drones were cheap, community-supported, and technically excellent — but built specifically to fly fast and crash safely. No payload concept. No structured data output. No GPS-assisted navigation. No documentation for anything beyond flying fast.

None of these options served the municipality, the community preparedness group, the university research team, or the volunteer SAR coordinator. The gap between "I can afford it but it is proprietary" and "it is open but I cannot afford it" was real and unfilled.

What libdrone actually is¶

libdrone is the EU civilian reference implementation for open-source aerial infrastructure. It is not a product you buy. It is a platform you build, understand, and own completely.

Designed in the Czech Republic. Licensed under CERN OHL-S v2 — the strongest open hardware copyleft licence available, developed at CERN for exactly this class of infrastructure. Legal entity in Estonia. All design files, every component choice, every engineering decision publicly documented and permanently preserved.

The hardware summary: a 330 mm quadrotor, 6-inch propellers, approximately 720 g all-up weight with battery. EASA Open A2 category — operable under EU drone law without an operator licence in most scenarios. 12–15 minutes endurance. The structural frame costs €16 in filament on any consumer FDM printer. Total build cost approximately €1,400 with goggles and a complete sensor payload.

The complete software and firmware stack is open source: Betaflight (GPL v3) on the flight controller, ExpressLRS (GPL v3) for the radio link, AM32 (MIT) for ESC firmware, EdgeTX (GPL v2) on the transmitter, FreeCAD (LGPL v2) for CAD. Every line of code is publicly readable and independently verifiable. No proprietary firmware in the critical path.

The satellite navigation uses Galileo and EGNOS — EU-origin infrastructure — in addition to GPS. This removes dependency on US military policy decisions. GPS Selective Availability demonstrated that this dependency is real; Galileo removes it.

The payload interface: why this is infrastructure, not a product¶

The feature that separates libdrone from any drone-with-sensors is the GX12-7 payload interface. Two aviation-grade connectors on the sealed top surface of every platform provide: regulated power, GPS coordinates at 10 Hz, bidirectional communications (UART and I2C), radio-controlled switching, and live overlay in the pilot's goggles.

Any sensor, camera, or instrument built to this interface works on any libdrone platform, regardless of version or manufacturer. An institution that invests in building an air quality payload can fly it on a Pro, a Bandit, or a future platform variant — without electrical rework, without a developer programme, without NDA. The interface specification is public.

→ gx12-connector-standard contains the full ICD.

This is the architectural distinction between a drone and a platform. A drone with a sensor is a single-purpose tool. A platform with a documented interface is infrastructure that compounds in value as more payloads are built for it.

Reference¶

Who is using libdrone and for what¶

Community preparedness groups deploying thermal and air quality payloads for flood route assessment, chemical incident monitoring, and night perimeter awareness. The drone that has been flown 50 times before the flood exists — the drone in a box does not. → civilian-preparedness, → resilience-use-cases

Municipal civil protection offices in Central Europe specifying a community-level aerial awareness capability at €1,400 rather than €5,000–50,000. The cost differential is not marginal — it is the difference between deployment and continued non-deployment. → sk-municipal-emergency-guide

University research groups using the payload interface to deploy novel sensors without custom drone development. The SEN66 air quality payload, the radiation survey payload, and the thermal imaging payload are documented and reproducible. → sk-university-programme-proposal, → payload-sdk

Civil defence and volunteer SAR groups in NATO-adjacent countries building autonomous situational awareness capability on the Bandit platform with ArduPilot, ATAK integration, and MAVLink telemetry. → bandit-variant, → iff-architecture

Cost comparison¶

Platform	Cost	Open source	Payload interface	Data sovereignty
DJI Mavic 3 Enterprise	€4,800	No	Proprietary SDK, NDA	No
Parrot Anafi USA	€7,000	Partial	Limited	Partial
Academic PX4 build	€10,000–50,000	Yes	Custom	Yes
libdrone Pro (full build)	~€1,400	Yes	Open standard	Yes
libdrone Core (training)	~€320	Yes	Open standard	Yes

The EU regulatory position¶

libdrone operates in EASA Open A2 category at standard configuration (~720 g). This means: no operator licence required in most EU member states for flights away from uninvolved people, no prior airspace notification required for most operations, no remote ID infrastructure dependency in Open category.

The CERN OHL-S v2 licence satisfies open science requirements for EU Horizon Europe and national research funding frameworks. The full FOSS stack satisfies EU Cyber Resilience Act auditability requirements. Galileo + EGNOS navigation satisfies EU strategic autonomy requirements where applicable.

→ easa-open-category, → foss-stack-libdrone, → regulatory-overview

The supply chain argument in one paragraph¶

Every libdrone component has a documented alternative. The BOM is public. The suppliers are documented. An institution can audit the supply chain independently without trusting the platform designer's claims. If AliExpress stops shipping, the alternative components are documented. If a specific ESC manufacturer discontinues a model, the compatible replacement is documented. No single company can strand a libdrone deployment.

→ bom-summary, → vendor-lock-in

Procedure¶

If you are evaluating libdrone for institutional procurement¶

Start here: → sk-platform-brief — 30 minutes, takes you from zero to a procurement recommendation.

Then: → resilience-use-cases for the 15 documented operational use cases.

Then: → bom-summary for the cost structure and supplier map.

Then: → sk-libdrone-eu-nato-bridge if your context involves NATO-adjacent governance or EU institutional documentation requirements.

If you found this article from a search and want to understand more¶

→ platform-overview is the correct next article. It covers the complete variant family, the payload interface in more depth, and the three gaps in the market that libdrone fills.

→ why-build-a-drone explains the engineering and preparedness case from first principles.

→ civilian-preparedness explains the specific use case for community and household preparedness deployment.

Rationale¶

Why this article exists¶

The people who need libdrone most are not searching for "libdrone." They are searching for "DJI alternative Europe," "open source drone EU," "drone data sovereignty," "affordable community preparedness drone," or "municipal drone program open source." This article is written for those searches. It names the problem they are experiencing before it names the platform that addresses it.

The goal is not to sell libdrone. It is to give an honest account of why DJI stopped being a default for EU institutional use, what the alternatives actually offer, and what libdrone is — so that anyone who lands here from a search can decide quickly and accurately whether this platform fits their context.