Vehicle tracking Drone Guide

Vehicle tracking is an increasingly important drone application for policing because vehicles can move quickly across complex urban, suburban and rural environments where fixed CCTV, ground patrols and helicopter support may not always provide continuous coverage. A drone can give officers a mobile aerial view of a vehicle, maintain situational awareness from a safer stand-off distance and help coordinate ground units without requiring them to remain immediately behind the target.

For policing, the strongest use case is not aggressive pursuit. It is observation, coordination and evidence gathering within lawful operational frameworks. A drone equipped with high-resolution RGB cameras, thermal imaging, optical zoom and AI-assisted target tracking can follow a vehicle through authorised areas while transmitting live video to a command centre or officers on the ground. The system can also help officers understand where the vehicle is heading, whether occupants have exited, whether other vehicles are involved and what risks exist around the surrounding road network.

The greatest value comes when drone data is integrated with police command systems, authorised CCTV, mapping and incident-management platforms. In that model, the drone becomes one mobile sensor within a larger operational picture rather than a standalone surveillance tool.

What Is Drone-Based Vehicle Tracking for Policing?

Drone-based vehicle tracking involves using an unmanned aircraft to maintain visual observation of a vehicle during an authorised police operation. The drone may be launched manually by a specialist unit or, in more advanced systems, from a remotely operated docking station positioned at a police facility or strategic location.

Once the target vehicle has been identified, the camera can remain pointed towards it while the aircraft follows from a safe and lawful position. AI may assist by placing a tracking box around the vehicle and maintaining the target within the frame as it moves.

Human officers remain responsible for operational decisions. The AI supports camera tracking and situational awareness but should not independently determine whether a vehicle is suspicious or what enforcement action should be taken.

Why Police Use Drones for Vehicle Tracking

Traditional vehicle pursuits can create substantial risks to officers, suspects and members of the public. Aerial observation can sometimes allow ground units to reduce their proximity while still maintaining awareness of the vehicle’s location.

A drone can provide an overhead view without requiring officers to follow directly behind the target at all times. This may help incident commanders coordinate road units more strategically and understand the wider situation before deciding how to proceed.

Drones can also reach areas that fixed CCTV does not cover and may be significantly faster and less expensive to deploy than a crewed helicopter in some situations.

Supporting Safer Pursuit Management

One of the most important policing applications is supporting pursuit management rather than simply increasing pursuit capability. If officers can continue observing the vehicle from the air, ground units may be able to maintain greater distance where operational procedures allow.

The drone can report location, direction of travel and changes in vehicle behaviour. This can help commanders make better-informed decisions about whether ground units should continue close observation, reposition or coordinate with other teams.

The objective is improved information and risk management, not encouraging more aggressive pursuit tactics.

AI Vehicle Detection

AI vehicle detection can identify cars, vans, trucks and motorcycles within the drone’s video feed. Once the relevant vehicle has been confirmed, the system can maintain a visual track even as the aircraft changes position.

The AI reduces the workload on the camera operator because they do not need to continuously adjust the gimbal manually. This allows the operator to focus more attention on the operational environment and communication with officers.

Human confirmation remains important, particularly where several similar vehicles are present.

AI Vehicle Classification

Computer vision can also classify broad vehicle categories such as passenger car, van, truck or motorcycle. This may provide additional context during an incident, particularly when the initial description is limited.

Colour and visible shape can sometimes assist identification, but these attributes should be treated cautiously because lighting, camera angle and distance can affect appearance.

Automated classification should therefore supplement confirmed police information rather than replace it.

Target Lock and Gimbal Tracking

A three-axis gimbal allows the camera to remain focused on the target independently from the drone’s direction of flight. The aircraft can move along a safer route while the camera continues looking sideways or backwards towards the vehicle.

Once AI target lock is established, the system continuously adjusts the gimbal to keep the vehicle centred. If the target approaches the limits of gimbal movement, the aircraft can reposition.

This creates smoother and more reliable observation than attempting to control every movement manually.

Optical Zoom

Optical zoom allows the drone to maintain greater distance while still capturing useful visual information. This can reduce the need for the aircraft to fly directly above a moving vehicle or congested road.

Zoom can help officers confirm the type of vehicle, visible damage or other exterior characteristics where resolution and legal authority permit. It can also provide useful imagery when the drone must remain outside a sensitive area.

Strong stabilization is important because high magnification makes aircraft movement more visible.

Thermal Vehicle Tracking

Thermal cameras can support police vehicle tracking at night or in areas with poor lighting. A recently operating vehicle may show clear thermal contrast because of the engine, exhaust, tyres and braking systems.

Thermal imagery can also help maintain observation if a vehicle leaves a lit road and enters a darker rural area or industrial site. It is especially useful when the drone needs to locate a vehicle that has stopped shortly before.

Thermal cameras generally provide much less identifying detail than RGB cameras, so dual-sensor payloads are more effective than relying on infrared alone.

Night Operations

Nighttime is one of the strongest operational environments for thermal police drones. A vehicle travelling on an unlit rural road may be difficult to follow visually, while its heat signature can remain prominent.

Low-light RGB cameras can then provide additional contextual detail once street lighting or other illumination becomes available.

Searchlights may be used where operationally appropriate, but they also reveal the drone’s presence and can affect the behaviour of people involved in the incident.

Vehicle Tracking After Loss of Ground Contact

A ground unit may temporarily lose visual contact because of road layout, traffic or operational safety decisions. An aerial drone may still be able to maintain observation.

The aircraft can provide updated location information to nearby units without requiring officers to immediately close the distance again.

This can be especially valuable where the operational objective is containment or coordinated interception rather than continuous close pursuit.

Urban Vehicle Tracking

Urban environments offer both advantages and challenges. Buildings, street lighting and road networks provide strong visual reference, but structures can block the drone’s line of sight and create complex airspace.

AI route prediction can use road geometry to estimate where the vehicle may reappear if it passes behind a building. Fixed authorised CCTV can also take over observation temporarily.

The most effective system therefore combines drones with other police sensing infrastructure rather than attempting to maintain one uninterrupted aerial camera view in every circumstance.

Rural Vehicle Tracking

Rural areas may have fewer fixed cameras and longer distances between police units, increasing the potential value of aerial observation. Thermal cameras can also perform strongly where background lighting is limited.

However, endurance becomes more important because the vehicle may travel farther. Cellular connectivity may also become weaker.

Long-range communications and safe return-energy management are therefore particularly important in rural operations.

Motorway and Highway Tracking

High-speed roads create another potential application because ground pursuit can involve considerable public risk. A drone may provide additional situational awareness from above where airspace and operating conditions allow.

The camera can observe which exits the vehicle takes, whether it stops or whether occupants leave the vehicle.

The drone should remain focused on observation rather than influencing the target’s driving behaviour.

Industrial and Port Areas

Police may also use drones during operations in ports, industrial estates or logistics areas. These environments can contain large internal road networks, warehouses and parked vehicles that make ground observation difficult.

Thermal imaging and elevated viewpoints can help maintain a target track through these areas.

Operational coordination with site security and other authorities may be important.

Vehicle Re-Identification

Re-identification becomes important if the target disappears temporarily and several similar vehicles are present when it reappears. AI can compare visible shape, colour, route and other characteristics to estimate which vehicle is most likely the original target.

Location continuity is particularly valuable. If the vehicle disappears behind a building and one similar car emerges seconds later from the expected road, that geographic context strengthens the track.

Human officers should still confirm identity where consequential actions depend on it.

Handling Multiple Vehicles

Some incidents may involve more than one vehicle. AI can maintain separate track IDs for multiple targets, but operational complexity increases quickly.

The system should make it clear which vehicle is the primary target and which are secondary observations. Confusing two vehicles could create serious consequences.

For this reason, human supervision and clear target confirmation remain essential.

Vehicle-to-Person Transition

A police vehicle-tracking operation may change when occupants stop and leave the vehicle. If legally authorised and operationally appropriate, the drone can transition from vehicle tracking to person tracking.

The aerial view can help officers understand where people moved relative to buildings, woodland, roads or other hazards.

AI can assist with target continuity, but human officers should remain responsible for deciding which person is relevant to the incident.

Thermal Person Tracking After Vehicle Exit

Thermal cameras can be particularly useful when occupants leave a vehicle at night and move into dark areas. The system may maintain awareness as the incident changes from vehicle observation to ground search.

This can support officers arriving on scene by providing a wider view of the surrounding environment.

The technology should remain integrated with established police search and command procedures.

CCTV Handover

A strong policing workflow can hand the target between drone cameras and authorised fixed CCTV. When the vehicle enters an area with excellent camera coverage, the drone may reposition or conserve battery.

If the vehicle leaves that camera network, the drone can resume primary observation.

This reduces unnecessary aircraft movement and helps maintain longer operational coverage.

Camera-to-Drone Handover

The reverse can also occur. Fixed CCTV or another police sensor may identify the vehicle first and send the location to the drone team.

The aircraft flies towards the reported position and attempts to visually acquire the target. Once confirmed, automated tracking begins.

This can shorten search time significantly.

Automatic Number Plate Recognition Integration

ANPR systems are generally better suited to fixed road positions than moving aerial platforms because they provide controlled angle, distance and lighting. Police vehicle tracking can therefore benefit from combining ANPR with drones rather than expecting the drone to perform reliable plate recognition at all times.

A roadside ANPR camera may identify the relevant vehicle, after which the drone provides broader movement tracking.

This division of roles makes the complete system more reliable.

Mapping and GIS Integration

The drone’s target track can be displayed on a digital map used by the police command centre. Officers can see roads, buildings, bridges and other geographic features around the vehicle.

The map can also show nearby units and authorised operational boundaries. This provides more useful context than video alone.

Geospatial tracking becomes particularly valuable when several teams are coordinating around the same incident.

Route Prediction

Road networks constrain where a vehicle can travel. Software can use map data, speed and direction to estimate likely next positions.

If visual contact is briefly lost, the drone can move towards the most likely reappearance point rather than remaining over the last known position.

These predictions should be treated as operational aids rather than certainty.

Live Coordinates

The system can estimate the target vehicle’s geographic position and transmit it to the command centre. Officers can then use this information to coordinate units without needing to interpret the drone video manually.

Position accuracy depends on drone GNSS, gimbal calibration, target range and viewing angle.

Laser rangefinding can improve target geolocation on some professional payloads.

Laser Rangefinder

A laser rangefinder can measure the distance between the drone and a selected vehicle or ground point. Combined with aircraft position and gimbal angles, this can improve geographic target location.

This may be useful in large open areas where accurate coordinates help ground units.

Any rangefinding system needs to be suitable for the operational environment and used in accordance with relevant requirements.

Drone-in-a-Box for Policing

Drone-in-a-Box systems could give police forces faster access to aerial observation because the aircraft remains charged and ready at a station or strategic location.

Instead of waiting for a specialist drone team to travel to the incident, a remotely supervised aircraft could potentially launch quickly within an approved operating area. Vehicle tracking would be only one of several missions, alongside missing-person searches, incident assessment and emergency response.

Such systems require strong operational governance, reliable communications and appropriate regulatory approvals.

Strategic Drone Stations

Police forces could position autonomous drone stations in selected areas according to operational demand. Urban stations may provide rapid coverage around city centres, while rural locations could support highways or remote communities.

The value depends heavily on launch response time and the legally approved operating radius.

A regional network could provide broader aerial coverage than relying on one central drone unit.

Remote Operations Centres

A central police drone operations centre could supervise several aircraft and stations. Trained personnel manage airspace, mission health and sensor use while local officers receive the resulting situational awareness.

This separates aircraft operation from the ground incident team and reduces the workload on officers at the scene.

AI can assist with routine target tracking, but operational control remains human.

BVLOS Operations

Vehicle tracking may require the drone to operate beyond the visual line of sight of a local pilot. This is one reason BVLOS capability could become important for police drone programmes.

BVLOS allows the aircraft to maintain coverage across larger geographic areas, but it introduces greater requirements around command-and-control links, airspace risk, contingency procedures and regulatory approval.

The operational case needs to be assessed carefully for each environment.

Multi-Drone Handover

Long incidents may exceed the endurance of one battery-powered drone. A second aircraft can launch and take over observation before the first returns.

The two drones briefly observe the same vehicle so the target can be confirmed before handover.

This approach can maintain continuous aerial awareness without pushing one aircraft beyond safe battery reserves.

Multi-Drone Operations

Large metropolitan areas may eventually use several police drones operating within defined zones. As a vehicle crosses from one area into another, observation responsibility can transfer to the nearest aircraft.

This reduces unnecessary long-range flight and improves response times.

Central coordination is essential to avoid airspace conflicts between the drones themselves.

Battery Endurance

Vehicle tracking creates unpredictable mission length. The drone should therefore maintain a protected reserve rather than simply remaining airborne until the battery becomes critically low.

The system can calculate required return energy continuously based on distance, wind and docking location.

Operational procedures should define when tracking must be handed over or terminated.

Return-to-Base Logic

A drone may be following a vehicle away from its launch location, meaning return energy increases continuously. Intelligent mission software should calculate when continued observation would compromise the safe return.

At that point, another aircraft may take over or the drone may return while fixed cameras and ground units continue observation.

Safety should remain independent from the urgency of the incident.

4G and 5G Communications

Cellular networks can support remote police drone operations by carrying telemetry or video where coverage is sufficiently reliable.

Private or priority communications networks may provide additional resilience for emergency services.

However, the aircraft should not rely on one communications path alone for safety-critical operation.

Lost-link contingency procedures remain essential.

Direct Radio Links

Direct RF links provide low-latency connectivity and may be useful where the drone remains relatively close to a police unit or station.

Urban buildings can block these links, while rural range may become the limiting factor.

Professional police systems may therefore use multiple communications technologies.

Satellite Communications

Satellite communications may support remote or rural operations where terrestrial connectivity is limited. The system can transmit essential telemetry and selected imagery while storing higher-quality footage onboard.

Latency and bandwidth need to be considered.

For many near-city police operations, terrestrial systems are likely to remain more practical.

Edge AI

Running vehicle detection and tracking onboard the drone reduces reliance on network bandwidth. The aircraft analyses the full-resolution video locally and transmits the relevant tracking information.

This also reduces latency, which can improve target tracking during rapid movement.

Edge AI is therefore particularly valuable for policing where response time matters.

Security and Encryption

Police drone video and target data can be highly sensitive. Command links, stored imagery and remote-access systems therefore require strong cybersecurity and encryption.

Access should be restricted according to role and operational need. Audit trails should show who viewed or exported data.

This becomes increasingly important as drone systems connect directly with wider police IT infrastructure.

Evidence Handling

Some drone footage may later become relevant to investigations or legal proceedings. If so, the original data should be preserved according to established evidence-management procedures.

Time, location and system metadata can help establish the context of the recording. AI annotations should remain traceable back to the original imagery rather than replacing it.

Evidence integrity should be designed into the system from the beginning.

Incident Recording

A police drone system can automatically preserve relevant video from before and after a tracking event. This creates a structured record showing when the target was acquired, where it travelled and how the incident developed.

The footage may be useful for debriefing, training or investigation where appropriate.

Retention should remain subject to applicable legal and organisational policies.

Human-in-the-Loop Control

AI can track a vehicle very effectively, but policing decisions should remain with trained officers. The system should not automatically conclude that a vehicle is committing an offence or direct enforcement action solely from computer vision.

Human operators confirm the target, interpret the wider context and decide how the information should be used.

This separation between automation and operational judgement is particularly important in policing.

Privacy and Proportionality

Vehicle tracking by police can involve significant privacy considerations because it may capture vehicles, people and properties unrelated to the incident. Operations should therefore be based on lawful authority, necessity and proportionality.

Camera operators can minimise unnecessary observation by maintaining appropriate zoom and framing around the operational target.

Data retention and access should also be limited according to the legitimate purpose of the operation.

Camera Geofencing

Camera geofencing can restrict the gimbal from pointing into areas that are not relevant to an authorised operation. This can help reduce unnecessary capture of residential or sensitive locations.

The restrictions can be defined geographically or according to mission type.

For emergency operations, authorised users may have different permissions from routine patrol missions.

Drone Geofencing

The aircraft can also operate within predefined flight boundaries. This helps prevent entry into prohibited airspace or areas outside the authorised police operation.

Dynamic geofencing could update as an incident moves while still respecting fixed airspace restrictions.

Strong geofencing becomes particularly important for remotely operated police drones.

Facial Recognition Considerations

Vehicle tracking does not inherently require facial recognition. In many operational situations, the objective is simply maintaining awareness of the vehicle and its location.

If a camera captures identifiable occupants, use of biometric identification should remain subject to the specific legal framework and police policy governing that technology.

Keeping vehicle tracking separate from unrelated biometric analysis can also reduce unnecessary data collection.

Public-Road Operations

Tracking vehicles on public roads raises greater legal, privacy and aviation considerations than monitoring a vehicle inside a controlled police or industrial site.

The drone must be operated under the applicable aviation framework while the policing activity itself also requires lawful authority.

Operational procedures should therefore clearly define when and how aerial vehicle tracking may be used.

Airspace Coordination

Police vehicle tracking can move quickly across different areas, which may bring the drone closer to airports, heliports or other restricted airspace.

The drone team needs real-time awareness of these constraints.

An operation may need to terminate or hand observation to other resources rather than allowing the aircraft to enter unsuitable airspace.

Helicopter Coordination

Police helicopters and other crewed aircraft may operate during the same incident. Drone and helicopter operations therefore require clear coordination.

A small unmanned aircraft should never create a conflict with crewed emergency aviation.

Where a helicopter arrives, established procedures may require the drone to leave the area or operate within a separately defined altitude and location.

Ground Unit Coordination

The drone’s greatest value comes when ground officers can use the information effectively. The operations centre can provide direction of travel, vehicle location and relevant environmental context.

Instead of every officer trying to watch the live video feed, a dedicated operator can communicate concise updates.

This helps avoid information overload during a fast-moving incident.

Tactical Situational Awareness

The aerial view may reveal information that is difficult to see from street level, such as roadblocks, queues, alternate routes or people leaving a vehicle.

This broader understanding can help commanders plan safer responses.

The drone should support tactical decision-making without attempting to automate those decisions.

Search After Vehicle Abandonment

If the vehicle is abandoned, the drone can help establish where occupants went immediately afterwards. Thermal and RGB cameras can search nearby open areas while ground teams respond.

The system can also provide a wider view of potential hazards around the abandoned vehicle.

This can be especially useful at night or in rural environments.

Vehicle Location in Car Parks

Large car parks can make it difficult to identify a vehicle once it stops among many similar vehicles. If the drone maintains continuous tracking during entry, it can record exactly where the target parked.

Thermal information may also help distinguish a recently operated vehicle from vehicles that have been stationary for longer.

Human officers still need to confirm vehicle identity before acting on the information.

Vehicle Tracking Through Woodland or Cover

Trees and covered structures can break visual contact. A drone may move to another angle or predict likely exit points based on the road network.

Thermal cameras do not allow reliable tracking through solid cover or dense structures, despite common misconceptions.

If the target is not visible, the system should treat it as lost rather than pretending it can continue seeing through obstacles.

Vehicle Tracking in Poor Weather

Rain, fog, wind and snow can all reduce drone availability and detection quality. Vehicle tracking should therefore remain one tool among several rather than the only method of maintaining observation.

Police forces need alternative procedures when weather prevents safe flight.

This may include ground units, CCTV or crewed aviation depending on the situation.

AI False Positives

Computer vision can misidentify objects, particularly when imagery is distant or poor. AI may also select the wrong vehicle after a crowded intersection or temporary occlusion.

The operator should therefore see the underlying video continuously and be able to correct the track immediately.

Consequential police action should never depend solely on an unverified automated target selection.

AI False Negatives

The AI may also lose the target because of glare, darkness, weather or visual obstruction. Confidence indicators can help operators understand when tracking quality is degrading.

The system should make uncertainty visible rather than hiding it.

A clearly reported lost track is safer than an incorrect track presented with false confidence.

Automated Reacquisition

When the vehicle disappears briefly, the system can search nearby roads for a likely match. It can use the previous speed, direction and appearance to narrow the candidates.

Once a likely vehicle is found, the operator confirms whether it is the correct target.

This combines automation with human verification.

Drone Search Patterns

If visual contact is lost completely, the drone can search likely roads or areas rather than flying randomly. GIS information and last known movement can define a structured search.

The aircraft may also climb to gain a broader view, subject to airspace and operating limits.

Search logic should remain conservative and transparent to operators.

Use After Vehicle Theft

Drones may assist in locating or observing a stolen vehicle when police already have lawful grounds and a relevant location. A drone could maintain visual contact while ground resources respond.

This may reduce the need for immediate close pursuit.

The aircraft should not be used as a general-purpose tool to track vehicles without appropriate legal basis.

Supporting Serious Incident Response

Vehicle tracking may also support wider serious-incident operations where the vehicle is one element of a developing event. The drone can provide commanders with an aerial view of both the vehicle and surrounding environment.

This can include understanding traffic conditions, identifying where occupants exited or observing the vehicle’s proximity to other people.

Human command remains central throughout.

Post-Incident Reconstruction

Recorded drone tracks and video can help reconstruct how an incident unfolded. Investigators may be able to review route, timing and vehicle movement alongside ground reports.

Where footage forms part of an investigation, evidence-handling procedures become important.

AI-generated tracks should be treated as analytical overlays on top of preserved original data.

Benefits of Police Vehicle Tracking Drones

The main benefit is mobile aerial situational awareness. A drone can follow a vehicle beyond the field of view of one fixed camera and give officers a broader understanding of what is happening around it.

It can also help reduce the need for close ground observation in some situations, provide night capability through thermal imaging and support coordinated response across several police units.

When integrated with CCTV, GIS and incident command systems, the drone becomes a highly flexible observation platform.

Faster Deployment Than Crewed Aviation

For some incidents, a locally based drone can be airborne significantly faster than a helicopter can reach the area. It can also operate at lower cost for routine observation.

Crewed aviation retains major advantages in range, endurance, payload and operational capability, so the two technologies should be considered complementary.

The best resource depends on the scale and seriousness of the incident.

Reduced Officer Exposure

Remote observation may reduce the need for officers to remain directly behind a vehicle simply to know where it is. This can lower exposure in some operations.

Ground units still need to handle the actual policing response.

The drone’s contribution is information and coordination.

Better Command Decisions

An incident commander who can see the vehicle, surrounding roads and nearby people simultaneously has a stronger operational picture.

This can support decisions about containment, unit positioning and whether continued observation remains appropriate.

The aerial view should therefore be shared in a way that assists rather than overwhelms command staff.

Challenges and Limitations

Police vehicle tracking with drones has significant limitations. Battery endurance, airspace constraints, communication coverage and weather can all interrupt operations. Buildings, tunnels and covered parking areas can break visual contact, while AI can confuse similar vehicles.

Legal and privacy requirements also matter greatly because public-road vehicle tracking can involve substantial observation of unrelated people and property.

Drones should therefore be treated as one component of a wider policing system rather than as a replacement for officers, fixed cameras or crewed aviation.

The Future of Vehicle Tracking for Policing

Police vehicle tracking is likely to become increasingly integrated with wider command-and-control systems. The future workflow will probably begin before the drone is airborne.

An authorised CCTV or ANPR system may identify the relevant vehicle and send its location to the closest available police drone. The aircraft then launches from a station or remote dock and travels towards the predicted route.

Once the operator confirms the target, AI handles routine gimbal tracking while the drone chooses safe observation positions based on roads, airspace and battery state. Fixed cameras can temporarily take over when they provide a better view, while the drone repositions.

If the target travels beyond one drone’s safe endurance or operating area, another authorised aircraft could take over. The system would preserve one continuous incident track even though several sensors contributed.

Thermal cameras, edge AI and improved communications will strengthen night and remote operations, while object-relative tracking and better target re-identification will reduce the chance of losing the vehicle.

The major transition will therefore be from manually piloting a police drone behind a vehicle towards integrated aerial incident tracking, where drones, authorised cameras, mapping and command systems work together to maintain situational awareness while officers remain responsible for every consequential policing decision.

Conclusion

Vehicle tracking is a valuable police drone application because it can provide continuous aerial situational awareness during incidents where ground officers or fixed cameras may otherwise lose sight of a vehicle.

High-resolution RGB cameras, optical zoom and thermal imaging allow the drone to observe vehicles during both daylight and nighttime conditions, while AI can reduce operator workload by maintaining the target within the camera frame. GIS and command-system integration then convert that visual observation into useful geographic information for officers on the ground.

The strongest operational value is supporting safer and better-informed incident management rather than simply extending high-speed pursuit. Aerial observation can sometimes allow officers to maintain awareness from greater distance, coordinate units more effectively and understand the wider environment around the vehicle.

Drone-in-a-Box systems and remote operations may make this capability available more rapidly in the future, while handover between drones, CCTV and other authorised sensors can provide longer and more resilient coverage.

Drones do not replace police judgement, legal authority, fixed surveillance systems or crewed aviation. AI should support tracking rather than independently decide who should be followed or what action should be taken.

For police forces developing drone programmes, vehicle tracking is therefore best understood as an aerial situational-awareness capability: a mobile camera and mapping platform that can help officers maintain a clearer picture of a moving incident while keeping human command, legality, proportionality and public safety at the centre of the operation.