David Hopkins, Director and Head of Operations and Delivery at Fire Hawk Service, will present this paper at the Australian OSINT Symposium in Sydney on Thursday 10 September 2026.
More platforms, no framework
Aerial collection is becoming a significant source of OSINT across many sectors, including emergency services, mining, critical infrastructure and environmental monitoring. However, more collection does not automatically mean better decisions. Left unstructured, it just as easily produces information overwhelm. In dynamic, high-stakes environments, aerial data only becomes decision advantage once an intelligence framework is applied to it. That framework means requirements are defined, legal and ethical parameters are established, collection is analysed against those requirements and parameters, and the result is developed into an actionable intelligence product for the right decision-makers.
Most organisations running drone programs in these environments are collecting large volumes of data without that framework. They often aren't even aware such a framework exists, let alone built one.
Rapid adoption
The capability and cost benefits of drones are substantial. A drone covers ground faster than a person on foot and sees what a person on the ground cannot, at a fraction of the cost of a manned aircraft. That combination of capability and cost is driving the rapid adoption of drones in Australia. As of May 2025, there were 38,874 licensed drone operators registered in Australia, more than the 30,845 conventional crewed pilots.[1] While much of that growth comes from agriculture and other automated applications outside the scope of this paper, it shows that drone platforms are now capable and affordable enough for large-scale adoption across sectors as varied as construction, mining, defence, logistics, and emergency services.[2]
Emergency services
The value of drones is most pronounced for emergency services. According to a 2023 sector study, emergency services stood to gain more from drone adoption than any other sector, with a 10 per cent net productivity increase worth an estimated $460 million to the national economy, and early bushfire detection alone offering $1 to $8 billion in potential savings.[3] The monitoring of fire-prone areas during dry lightning storms, detecting strikes and addressing hot spots before they develop into major fires, is a capability with the potential to save lives and prevent billions of dollars in property damage.[4]
The benefits of drone-enabled aerial collection extend well beyond the fireground. CASA has been working to assist the rollout of drone-delivered capabilities across emergency services.[5] Western Australian police have been trained to fly drones beyond visual line of sight and close to people, a capability used inside buildings during an armed siege to reduce risk to officers, and CASA is now looking at extending that authorisation to other agencies. During Tropical Cyclone Alfred and the recent NSW floods, drones located people trapped by rising water, delivered hay to isolated farmers and urgent supplies to housebound patients, and lit the way for a night-time rescue. Search and rescue, maritime patrol and surf lifesaving all draw on the same underlying capability, covering large or hazardous areas quickly, spotting what a person on the ground or in a boat cannot, and doing it without putting a crew at risk.
Cross-sector value
In critical infrastructure, drone inspection is replacing manual survey methods that are slower, more expensive and often more dangerous. Instead of sending a person up a bridge tower or along a pipeline corridor, a drone can now do the same work from the air. Mining operations use aerial collection for stockpile volume measurements, tailings monitoring, and site safety, offering faster coverage, safer access to hazardous terrain, and more consistent, repeatable data over time. Agriculture has adopted drones for crop monitoring, spreading, and spraying, delivering real efficiency and yield gains. Environmental monitoring programs can use drones to track change across large or remote areas that would otherwise require far more time and cost to survey on the ground.
Decision advantage
Regardless of the sector, drones can deliver substantial efficiency, accuracy and safety gains. Turning those gains into decision advantage, meaning faster, more accurate, and more confident decisions, requires more than platforms and pilots. It needs a well-designed intelligence framework, which:
- Defines what's needed before collection starts
- Governs how the data is handled once it's captured, including legal authority, retention, and access
- Shapes the product around who needs to receive it – an incident commander making a decision in real time needs something different from an investigator building a case after the fact.
If any one of those parts is skipped, the volume of data an organisation can now collect becomes a liability rather than an asset, unusable, ungoverned, or delivered to someone who can't act on it. Get all three right, and the same data becomes decision advantage.
Looking through a straw
Flying a drone without an intelligence framework is like looking at the earth through a straw.
The decision-maker can see whatever is directly in front of them with extraordinary clarity, high resolution, and a real-time feed from multiple sensors. But their field of understanding is narrow, and their context is thin. They have no systematic way to connect what they are seeing to the decisions being made, or to any legal or ethical parameters the data is subject to. The straw can be moved, and more data can be collected. However, without an intelligence framework, none of it is measured against requirements, data caveats are unknown, and the unstructured flow of data becomes overwhelming and unusable for decision-makers.
A 2024 field trial with firefighters in Switzerland found that information overload is a primary challenge in autonomous drone operations.[6] Delivering large amounts of information at once can overwhelm a commander’s capacity to absorb it, leading to slower or less accurate decisions.[7] Firefighters in that trial identified solutions including delivering the data in a form that can be used by operators, mirroring the purpose of the intelligence framework.
Drone-sourced aerial intelligence is part of a wider OSINT stack, and it supports better decisions in the ops room, in assessments and in reports, when it's carefully integrated alongside other sources. In other words, the straw becomes useful when it is part of an intelligence framework.
The intelligence framework
The organisations that get full value from their drone investment treat it as part of their OSINT stack and operate within a well-designed intelligence framework. That starts with thinking about requirements, parameters and audience before the platform itself.
This approach is known as Precision Aerial Intelligence. It treats every flight as a response to a defined requirement. Collection is directed at that requirement, analysed against it, and delivered as a product shaped for whoever has to act on it. The framework is in place from the outset, before a drone program even begins.
Building that framework starts with the decision environment. What is the organisation trying to manage, protect, detect, or understand? What decisions need to be made? Once this is known, further questions need to be asked: what information is needed to fulfil the identified requirements, what information will impact decisions, and what information is already available from other sources? Where are the gaps, and is aerial collection the right way to fill them, within whatever legal and ethical parameters apply to what's being collected? Only once those questions have answers does it make sense to start a flight plan.
Drone-enabled aerial collection is part of the OSINT stack, more efficient, cheaper and safer than manned aircraft, and it fills specific gaps that other sources can't. Satellite imagery gives broad coverage, but for civilian and commercial organisations, not the persistence or real-time response a drone can offer on demand. Ground sensors give point-specific data but no spatial context. Human sources and public reporting give intent and indicators but not physical ground truth. Aerial collection earns its place in the stack by providing persistence, spatial context, and physical ground truth, all in real time and directed against a specific requirement.
Once those requirements are defined, collection planning works backward from them. This includes identifying what to collect, where, when, with what sensors, and to what standard. The drone is only assigned once those requirements are met, even when the platform itself is fixed, owned outright, or the only option available.
The difference between a drone program and an intelligence capability is whether requirements drive collection, or collection happens because the platform is there. The next section shows what this distinction looks like in practice, based on my own operational experience across two very different firegrounds.
Decision advantage in practice
Every fire incident draws on more than one source of information. The ops room works from radio traffic, resource tracking, weather forecasts, and systems like Athena that model fire behaviour and ingest social media reporting.[8] An incident controller on the ground adds direct observation, communication with crews on scene, and local knowledge of the terrain. A drone pilot contributes something different again: real-time aerial vision that none of those other sources can provide on demand.
Tomboye Fire, 2024
In 2024, I responded to the Tomboye fire near Bungendore, NSW as a drone pilot, with the NSW Rural Fire Service. The incident controller (IC) asked for help, and that is the entirety of the tasking I received. There were no defined collection requirements and no situation report (sitrep) provided.
My background in intelligence, combined with fireground experience, meant I knew what to ask before launching. I spent a few minutes working through the questions any collection discipline demands. The most pressing was what decision did the IC need to make right now? They needed to know where the fire edge was, where the hot spots were, what terrain features were relevant to fire behaviour and asset protection, and where their resources sat relative to the threats. Those needs became the collection requirements I used to develop the flight plan.
Four flights followed. The first oriented to the fire and mapped the boundary using FireMapper InFlight[9], feeding directly into the Athena system and Hazards Near Me.[10] The second and third refined the fire line, giving the IC a direct look at the western edge. The fourth targeted candling trees, smokers, and hot spots using thermal imaging, with identified areas of increasing activity uploaded back into FireMapper. In less than an hour of flying, the IC had a spatial picture of their fire, assets, and threats that would have taken significantly more time and posed greater risk to build any other way.
Using intelligence discipline turned a vague request for help into decision advantage, with a defined set of answers the IC could act on immediately.
Mayfield Road Fire, 2026
Eighteen months later, the same approach was applied at a different scale during the Mayfield Road Fire near Mulloon, NSW, in January 2026. The fire covered 124 hectares, burned under Extreme fire danger conditions, had a perimeter of 8.5 kilometres, and ran for four days and nineteen hours, with a full Incident Management Team (IMT) activated.
I was listed in the Incident Action Plan (IAP) as a named aviation resource alongside rotary wing aircraft, heavy plant, and sixteen ground appliances. The drone was planned into the incident structure before operations commenced, with formally defined tasking:
- investigate a structure on the western edge,
- map the active fire line on the NW perimeter where ground visibility was limited, and
- fly the full perimeter to update fire extent as time permitted.
The requirements were defined before the aircraft left the ground, and collection was directed against those requirements from the outset.
The output fed directly into the IMT's common operating picture via FireMapper. Post-incident accuracy checking showed my fire line mapping was near-identical to an FB100 aerial scan conducted the following morning, with 10-20 metre variation attributed to the difference between tracked shapes and manually annotated lines. The collection was accurate and integrated, providing the IMT with a decision advantage and critical information faster than any other method could have.
The after-action review (AAR), approved by the Lake George District Manager in June 2026, formally endorsed the continuation and expansion of drone operations. It noted that RPAS operations significantly enhanced situational awareness, particularly in inaccessible areas, and recommended integrating RPAS into incident planning and IMT processes as standard practice.
The AAR also flagged a connectivity problem. On one sortie, I had to leave the fireground to upload data, so the collection was accurate but reached the IMT later than it should have. Part of building an intelligence framework is interrogating how a product will reach the decision-maker. It doesn't remove every operational risk; connectivity will always be a variable in the field, but it means dissemination is considered part of the tasking rather than discovered as a problem after the fact.
What a good framework looks like
An organisation that has built drone-enabled aerial collection into its OSINT stack properly incorporates its drone program into the overarching intelligence framework. The drone itself is one collection asset among several, tasked against defined requirements. Its output enters a processing and analysis workflow before it reaches the decision-maker in a form they can act on, while it is still useful.
Collection planning happens before flight planning, so the pilot understands what they are collecting for and why. The raw data is processed, contextualised, and communicated in a form that meets the requirements it was tasked against. The decision-maker gets what they need to decide faster and with more confidence than they could have otherwise.
For the OSINT community
Drone-enabled and intelligence-led aerial collection, what I call Precision Aerial Intelligence, is fast, flexible, and far cheaper than manned aircraft, and it can be tasked and repositioned in real time in a way no other source in the OSINT stack can match. However, those benefits only become decision advantage once they are integrated into an intelligence framework. This framework defines requirements before the flight, accounts for the legal and ethical parameters to which the data is subject, and shapes the resulting product around the decision-maker who has to act on it. Building that framework and integrating Precision Aerial Intelligence into the OSINT stack turns a capable platform into a genuine source of decision advantage.
References
Li, M., Katsiuba, D., Dolata, M., & Schwabe, G. (2024). Firefighters' perceptions on collaboration and interaction with autonomous drones: Results of a field trial. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (Article 1). Association for Computing Machinery. https://doi.org/10.1145/3613904.3642061
Mason, R. (2023). New South Wales Rural Fire Service – Setting the stage. AerialFire Magazine. https://aerialfiremag.com/2023/07/03/new-south-wales-rural-fire-service-setting-the-stage/
NSW Rural Fire Service. (n.d.). FireMapper [Operational guideline]. https://www.rfs.nsw.gov.au/resources/publications/doctrine/technical/firemapper
NSW Rural Fire Service, South Eastern Area Command. (2026). Mayfield Rd Fire (Lake George) after action review (Insight 7.12) [Unpublished internal report].
Spence, P. (2025). RPAS Australian Skies Conference 2025, Civil Aviation Safety Authority, media release/speech, 2025. https://www.casa.gov.au/about-us/news-media-releases-and-speeches/rpas-australian-skies-conference-2025
Toll Uncrewed Systems. (2026). Complete Australian Drone Licence Guide 2026, www.tolluncrewedsystems.com/drone-licence/australian-drone-licence-guide
Wiedemann, M., Vij, A., Banerjee, R., O'Connor, A., Soetanto, D., Ardeshiri, A., Anilan, V., Wittwer, G., & Sheard, N. (2023). Validating the benefits of increased drone uptake for Australia: Geographic, demographic and social insights. University of South Australia, in partnership with iMOVE Cooperative Research Centre, for the Department of Infrastructure, Transport, Regional Development, Communications and the Arts. https://www.drones.gov.au/sites/default/files/documents/validating-the-benefits-of-increased-drone-uptake-for-australia-final-report.pdf
[1] Spence, P. (2025)
[2] Toll Uncrewed Systems. (2026)
[3] Wiedemann et al. (2023)
[4] Spence, P. (2025)
[5] ibid
[6] Li et al. (2024)
[7] ibid
[8] Athena is an AI-assisted bushfire intelligence platform developed by NSW RFS with Kablamo, combining fire scan and GIS data, ignition points, weather and social media monitoring into a single operational picture used for fire prediction and incident coordination. Reference: Mason (2023)
[9] FireMapper is the NSW RFS's in-field mapping application, used to share mapping information, photos and resource locations, and to support incident management, with the Incident Controller responsible for keeping the incident map updated. Reference: NSW Rural Fire Service. (n.d.)
[10] Hazards Near Me is the NSW Government's public-facing app (formerly Fires Near Me) providing real-time bushfire and other hazard information to the community, developed jointly by the Department of Customer Service, the SES and NSW RFS.
AI Use Disclosure
This paper was developed with AI assistance for research, structuring, and editing for clarity. All ideas, arguments, and conclusions are those of the author. Final responsibility rests with Fire Hawk Services, in accordance with Hire Hawk Holding's AI Use Policy.