以無人機建立即時自動化橋梁裂縫影像辨識系統
Other Title
DEVELOPING A REAL-TIME AUTOMATED BRIDGE CRACK DETECTION SYSTEM USING UNMANNED AERIAL VEHICLES (UAVS)
Journal
中國土木水利工程學刊
Journal Volume
37
Journal Issue
4
Start Page
269
End Page
277
ISSN
1015-5856
Date Issued
2025-06
Author(s)
Abstract
建築與橋梁的生命週期中，使用階段佔大部分時間與資源，特別是維護與檢修。傳統檢測依賴人工目視，需投入大量人力與時間，且存在安全風險。隨著人工智慧的發展，深度學習與無人機技術結合為基礎設施監測提供新方法。無人機具備高機動性，可克服傳統檢測的限制。然而，多數研究僅將無人機作為影像工具，缺乏即時處理能力。本研究整合機器人系統、無人機與影像辨識技術，建立即時橋梁裂縫檢測系統，可即時處理影像、辨識裂縫，並儲存結果與發送警示，提高監測效率與安全性，並適用於災後搜救與災害評估，提升基礎設施維護與應變能力。
In the life cycle of buildings, the usage phase accounts for majority of time and resource consumption, particularly in terms of maintenance and regular inspections. Traditional inspection methods primarily rely on manual visual assessments, which require substantial manpower, are time-consuming, and expose inspectors to potential safety risks, especially when working at heights or in hazardous environments. As artificial intelligence technology continues to advance, the integration of deep learning with drone technology has introduced innovative solutions for infrastructure monitoring. Drones, with their high mobility and ability to cover large areas efficiently, can effectively address the limitations of conventional inspection techniques. However, many existing studies have primarily utilized drones as tools for capturing images, requiring subsequent manual or semi-automated processing, which reduces real-time efficiency. To address this gap, this study integrates robotic systems, unmanned aerial vehicles (UAVs), and image recognition technology to develop a real-time bridge crack detection system. This system is capable of processing images in real time, identifying cracks using AI-driven algorithms, storing results systematically, and sending immediate alerts when structural issues are detected. By enhancing both the efficiency and safety of infrastructure monitoring, this system offers significant advantages over traditional inspection methods. Furthermore, its applications extend beyond routine maintenance to critical areas such as post-disaster search and rescue operations, as well as rapid disaster assessment. By leveraging AI and UAV technologies, this approach strengthens the overall resilience of infrastructure management, improves emergency response capabilities, and lays the groundwork for a more intelligent and automated future in structural health monitoring.
Subjects
深度學習
電腦視覺
無人機
機器人作業系統
deep learning
computer vision
UAV
ROS
Type
journal article