The Challenge: Detecting and Tracking Unauthorized Drones Over a Sensitive Facility

A critical infrastructure facility operating within a restricted airspace zone faced a growing threat from unauthorized drone incursions that existing perimeter security systems were not designed to address. Ground-level PTZ cameras and fence-line sensors provided effective perimeter security against personnel intrusion but offered no detection capability for aerial threats approaching from any altitude above 15 meters — effectively the entire operational envelope of commercial and modified consumer drones.

Incidents had included reconnaissance-type flight patterns over restricted areas, payload delivery attempts to secure zones, and signal interference consistent with drone-mounted electronic countermeasures. The facility's security team had no reliable method of detecting incursions until the drone was visible to on-site personnel, by which point meaningful response was no longer possible.

The requirement covered three distinct functions: detection of drone-sized aerial targets at sufficient range to enable response before the target reached sensitive areas; optical identification to confirm target type and flight characteristics; and real-time tracking to support response team coordination. The system also needed to integrate with the facility's existing security platform without replacing its ground-level camera infrastructure.

Why Standard PTZ Cameras Alone Were Not Sufficient

• Standard ground-level PTZ cameras — Optimized for horizontal field-of-view coverage, standard PTZ cameras lack the upward pan-tilt range and detection sensitivity needed for aerial target acquisition. Detection of a 0.5-meter drone at 300 meters requires a combination of thermal sensitivity and optical magnification that exceeds standard specifications.
• Radar-only systems — Radar detection provides range and bearing data but no optical confirmation of target type, generating high false-positive rates from birds and weather clutter in complex environments.
• Fixed thermal cameras — Effective for detection but limited to fixed fields of view, requiring the target to pass through a specific detection zone rather than enabling tracking across the full facility airspace.

Solution Architecture: Dual-Sensor EO/IR PTZ + AI Analytics Integration

The counter-UAS configuration deployed the FTD-PTZD Spectrum Dual-Sensor Network Positioning System at four elevated nodes positioned to provide overlapping spherical coverage of the facility airspace. The dual-sensor configuration — thermal imager for initial detection and optical zoom for identification — was integrated with the facility's FTD-16CH AI Analytics Server running a drone detection model trained on thermal and optical signatures of commercial drone platforms.

The AI analytics layer processed thermal imagery in real time, triggering automatic camera slew-and-track when signatures consistent with a drone heat profile were detected. The system's 36x optical zoom then acquired the target for visual confirmation, with the tracking algorithm maintaining lock through the target's flight path across the facility airspace — a capability that manual operator response could not replicate at the speeds and altitudes involved.

Integration with the facility's existing security platform used the FTD-PTZD's ONVIF output for video feeds and SDK-level integration for alarm relay, allowing drone detection alerts to appear within the existing security management interface alongside ground-level camera feeds and access control events.

Technical Specifications

Deployment Details

The four camera nodes were positioned at the corners of the facility's security perimeter at heights of 12–15 meters, providing spherical airspace coverage with overlapping detection zones at all approach angles. Elevated mounting used SUS304 stainless steel pole mount brackets on dedicated masts separate from the facility's ground-level camera infrastructure, preventing vibration from HVAC and industrial equipment from affecting the tracking precision of the aerial monitoring system.

The AI analytics server was deployed in the facility's existing server room, processing video feeds from all four cameras simultaneously. The drone detection model was pre-trained on a dataset of commercial drone thermal signatures and fine-tuned during a two-week commissioning period using controlled test flights to optimize detection thresholds for the specific thermal background conditions at the facility.

False positive management — distinguishing drones from birds, which produce similar thermal signatures at detection range — was addressed through a combination of target size estimation, flight path analysis, and optical zoom confirmation triggered automatically when a detection event occurred. The system reduced false positive rates to under 2 per 24-hour period during the operational period.

Results After 60 Days of Operation

• 11 unauthorized drone incursions detected during the 60-day period — none had been detected by the facility's previous ground-level security system during an equivalent prior period.
• Average detection-to-tracking time: 4.2 seconds from initial thermal detection to optical lock, enabling security response team notification before the target reached the inner security perimeter in all detected cases.
• False positive rate: 1.8 per 24-hour period, below the 2.0 target threshold and significantly lower than the 8–12 false positives per day recorded during the initial pre-tuning phase.
• Operator workload for airspace monitoring: reduced to near-zero during non-alert periods, with the AI system handling continuous airspace monitoring autonomously and alerting operators only when confirmed detections occurred.

Frequently Asked Questions

What detection range is achievable for commercial drones using thermal imaging?
Detection range for a 0.5-meter consumer drone depends primarily on the thermal sensor's sensitivity (NETD rating) and the temperature differential between the drone's motor heat signature and the ambient background. In typical conditions, a high-sensitivity thermal imager with NETD below 50mK can detect a commercial drone at 600–800 meters. Smaller drones or those with low-heat-output motors reduce this range; larger commercial platforms with multiple high-power motors extend it.

How does AI analytics distinguish drones from birds at thermal detection range?
At detection range, thermal signatures of drones and birds are similar in size and heat output. AI classification uses a combination of factors: target size estimation from pixel count at known magnification, flight path characteristics (drones maintain more consistent altitude and speed than birds), periodic motion analysis (bird wing-beat patterns versus drone motor vibration), and optical zoom confirmation when initial detection occurs. Multi-factor classification reduces false positive rates to levels suitable for operational security monitoring.

Can existing PTZ camera infrastructure be extended to include drone detection?
Existing PTZ cameras can contribute to a drone detection system if they have sufficient tilt range to cover the airspace above the facility and adequate optical zoom for identification. However, detection sensitivity for small targets at range typically requires purpose-specified dual-sensor EO/IR systems. The most effective architectures layer drone-specific detection cameras over existing ground-level PTZ infrastructure rather than replacing it.

What integration is available between drone detection systems and response coordination platforms?
ONVIF-compliant video feeds integrate with standard VMS platforms for recording and operator display. SDK-level integration enables alarm relay to security management systems, access control platforms, and communication systems for response team notification. For facilities requiring integration with air traffic control or regulatory reporting systems, custom data output formats are available via the analytics server API.

Related Products

• FTD-PTZD Dual-Sensor EO/IR Positioning System
• FTD-16CH AI Analytics Server
• IRW2D Enterprise Thermal EO/IR PTZ Camera
• SUS304 Pole Mount Bracket