Successful cases

To promote the development of low-altitude economy, the Government launched the Regulatory Sandbox pilot projects in early 2025. The Automatic Drone Patrol and Monitoring System, jointly developed by the EMSD and the Correctional Services Department (CSD), was selected as one of the first batch of pilot projects. Equipped with an automated drone dock, the system enables the drone to patrol the prison beyond the operator’s visual line of sight, effectively enhancing the operational efficiency of prisons.

EMSD’s First Beyond Visual Line of Sight Flight Project

The drone system is currently in operation at Tong Fuk Correctional Institution for fence and rooftop patrols. Its major breakthrough lies in achieving beyond visual line of sight (BVLOS) flight, which allows the drone to operate outside the operator’s visible range and thereby expands patrol coverage. Under the existing requirements of the Civil Aviation Department (CAD), drones must remain within the visual range of both operators and observers, thus limiting their scope of operation. Nevertheless, within the Regulatory Sandbox framework, the Government has, for the first time, relaxed this restriction and permitted BVLOS flight tests. To ensure flight safety, the drone must be equipped with additional sensors for automatic obstacle avoidance and precise positioning. The system passed compliance checks and completed more than 20 trial flights. Ultimately, it received approval from the CAD in October this year to officially conduct BVLOS flight missions.

Automated Drone Dock

The EMSD and the CSD launched a pilot scheme at Cape Collinson Correctional Institution in 2020 to test drone patrols of prison facilities. Two years later, the trial continued at Tong Fuk Correctional Institution. In 2024, the departments further enhanced the system by introducing an automated drone dock. The drone can follow preset routes to perform patrol missions and, upon completion, return to the dock for automatic charging, all without manual intervention. With its waterproof, dustproof and temperature adjustment features, the dock ensures that the drone inside remains undamaged even during periods when a Hurricane Signal No. 10 or Black Rainstorm Warning Signal is in force.

Artificial Intelligence Image Analysis

Through the fleet management system, CSD staff can set flight times and routes, and monitor the drone remotely during its operation. All flight data and images are automatically recorded. The system also employs artificial intelligence (AI) to analyse images in real time for automatic detection of fence damages or rooftop objects, and automatically generates reports after each flight, significantly enhancing patrol efficiency and accuracy.

Saving Patrol Manpower and Promoting Drone Development

Currently, the CSD deploys officers for daily foot patrols along rugged paths, during which they may encounter wild animals. Drone patrols not only save manpower and time, but also safeguard staff safety. Moving forward, the CSD will gradually replace part of the routine manual patrols with drone patrols according to operational needs and explore extending of the system to other suitable correctional institutions.

Given the capacity of drones to transcend geographical barriers, the system is also applicable for patrolling remote areas such as country parks. Furthermore, data obtained from the BVLOS flight tests will provide valuable reference for future legislative amendments by the Government. 

Apart from prison patrol, we are also actively expanding the applications of drones, including inspecting cemeteries, examining masts and towers, supporting outdoor rescue operations, and assisting in crowd management during large-scale events. The EMSD will continue to uphold the principles of the third Five-year Strategic Plan of the Electrical and Mechanical Services Trading Fund (EMSTF), working closely with clients to implement innovative solutions and applying technologies such as drones and AI to deliver customer-oriented, innovative services.