2020-06-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/712651摘要：大型建物與公共工程多為鋼筋混凝土結構，而影響鋼筋混凝土結構安全品質的關鍵則是鋼筋組立的施工品質。在工地現場進行澆注混凝土前，查驗鋼筋框架是否符合圖說是非常繁瑣但必要的步驟。大型建物與公共工程的施工現場往往相當複雜且具高風險，如深基礎開挖可達數十公尺深，高架橋墩柱可達數十公尺高，查驗人員處於此工作環境，查驗工作需顧慮施工環境安全。整體而言，對於鋼筋組立之查驗作業，查驗人員要面對鋼筋數量大、樣式多、量測不易、時效、安全考量等挑戰，但又需確保鋼筋組立施工品質的挑戰。本計畫的目的是研發以影像量測與人工智慧技術為基礎，有效輔助工地鋼筋查驗的軟、硬體技術、工具與作業模式。本計畫將分兩年執行，主題為以人工智慧輔助工地鋼筋工程查驗，第一年以資料蒐集、深度學習、模型訓練與立體影像量測為核心，已經建立完整查驗模組以利全面性鋼筋查驗和查驗結果紀錄，研發的技術已應用於較為單純的施工場地進行驗證。第二年將導入無人飛行載具進行複雜工地場域的資料搜集，並拓展第一年所研發的軟硬體技術，進行特殊場域進行資料蒐集與遠端輔助鋼筋查驗流程，減少作業人員處於高風險工作環境。期望將此輔助工地鋼筋查驗的技術應用至更多待檢核的鋼筋工程施工環境，協助查驗人員完成例行性的查核作業，達到品質查驗質、量的提升。<br> Abstract: Large-scale infrastructures are mostly reinforced concrete (RC) structures, andthe key to ensuring the safety of RC structures is the quality of reinforcementinspection. Before pouring concrete on the site, it is a very cumbersome butimportant step to inspect whether the reinforcements to design. In addition, theconstruction site is often complicated, for example, deep foundation excavationcan reach tens of meters deep, and the viaduct pier can reach tens of metershigh. The inspection personnel are exposed in a high-risk environment. Overallspeaking, for reinforcement inspection, the inspectors have to face thechallenges of large number of steel bars, many styles, difficult measurement,timeliness, safety considerations, etc. The objective of this study is to developadvanced technology and tools based on image measurement and artificialintelligence technology to effectively assist the on-site reinforcement inspection.This is a two-year project. The theme is to use the artificial intelligence to assistthe site reinforcement inspection. In the first year, the core was the developmentof the data collection, machine learning, data model training and stereo imagemeasurement. Complete reinforcement inspection modules were developed tocarry out the full inspection on reinforcement and the recording of the inspectionresult. The developed technology was also applied to the relatively simpleconstruction site. In the second year, unmanned aerial vehicles will be utilized tocollect data from complex sites to avoid putting inspectors in danger. Thesoftware and hardware technologies developed in the first year will also beexpanded. As a result, the proposed system is expected to be implemented inmore construction sites to assist the inspection of the reinforcement. Moreover, itis anticipated to improve the quantity and quality of the routine practice.鋼筋現地查驗工地人工智慧機器學習深度學習電腦視覺立體影像量測工地安全無人飛行載具Reinforcement Site InspectionConstruction SiteArtificial IntelligenceMachineLearningDeep LearningComputer VisionStereo Image MeasurementSafetyin Construction SiteUnmanned Ariel VehicleUAV