AI for Roadside Signboards Detection, Recognition & Identification, and Structural Analysis

I&T Wish AI for Roadside Signboards Detection, Recognition & Identification, and Structural Analysis
(REF: W-0537)
Matched I&T Solution
Trial Project
Summary and Challenges
  • Signboard control in existing buildings is crucial for public safety especially for pedestrians and building occupants. A patrol vehicle would be assigned for the inspection work on the signboards. By leveraging Artificial Intelligence (AI) for analyzing the roadside images collected via a camera setup on the patrol vehicle, signboards shown should be detected and recognized for further comparison with the same signboard detected from previous patrolling operations.
  • Further, by checking on the structural frames of the signboard, falling risks shall be identified, evaluated and explained. A user interface shall be developed for the users to review the findings and to proceed with required administrative workflows and checks.
Expected Outcome
  • The System shall be able to achieve the following:
    • Identifying and segmenting signboards from a wide-view of roadside image
    • Recognize and identify the signboards and associated frames and group them according in timeline on its changes through every operation
    • Perform structural analysis on signboard framework and identify its defects with sufficient explanation (i.e. explainable AI), defects including rusty, deformation, missing display, and more should be identified
    • Develop user interface for effective administrative workflows and process
    • Develop user interface for effective signboards lookups and checking with operations log
Expected Trial Duration 8-month
Contact Information
I&T Wish Proposer:Buildings Department (BD)
Contact Person:LEE Nicholas
Position:Electronics Engineer / Artificial Intelligence
Tel:3167 2896
Email: nicholaslee@emsd.gov.hk
Upload Date 2024-07-15
Closing Date 2024-07-29
Propose I&T Solution Closed