Development of Artificial Intelligence Powered (AI-powered) Video Analytic Platform for Drone Inspection

I&T Wish Development of Artificial Intelligence Powered (AI-powered) Video Analytic Platform for Drone Inspection
(REF: W-0593)
Matched I&T Solution
Trial Project
Summary and Challenges
  • Drone-based inspection requires an integrated platform that can seamlessly connect with flight control systems to extract video and image data. This platform should host a variety of AI-driven image analysis models, enabling automated detection, classification, and anomaly recognition. The results should be presented through interactive dashboards and comprehensive reporting tools, supporting decision-making and operational efficiency.
  • Key challenges include ensuring reliable data integration across diverse drone flight control platforms, managing large volumes of high-resolution imagery, and maintaining accuracy and adaptability of AI models for different inspection scenarios. Additional hurdles involve building scalable dashboards that can deliver actionable insights, and establishing standardized reporting mechanisms that meet regulatory and operational requirements.
Expected Outcome
  • This AI-powered Video Analytic Platform shall include the following:
    • Deployable both on cloud (via monthly subscription) and on-premise environments
    • Provides multi-brand drone flight control integration points ensuring broad compatibility
    • Enables drone flight control and scheduling for coordinated operations
    • Delivers sufficient GPU and CPU computing power for intensive workloads
    • Supports 5G and other communication technologies for real-time connectivity
    • Hosts a reinforced AI algorithm repository for scalable model management
    • Offers multiple AI algorithms, e.g. waste monitoring
    • Allows integration of third-party AI algorithms for extended functionality
    • Includes a dashboard platform for data visualization and operational insights
    • Provides reporting and alarm/alert mechanisms for timely decision-making
    • Supports data labeling to improve model accuracy and continuous learning
Expected Trial Duration 9-month
Contact Information
I&T Wish Proposer:Electrical and Mechanical Services Department (EMSD)
Contact Person:LEUNG Sze Ming
Position:Project Officer/Innovation/40
Tel:9151 2597
Email: leungsm@emsd.gov.hk
Upload Date 2025-11-26
Closing Date 2025-12-10
Propose I&T Solution Submit Proposal