Skip navigation

DSpace JSPUI

DSpace preserves and enables easy and open access to all types of digital content including text, images, moving images, mpegs and data sets

Learn More
DSpace logo
English
中文
  • Browse
    • Communities
      & Collections
    • Publication Year
    • Author
    • Title
    • Subject
    • Advisor
  • Search TDR
  • Rights Q&A
    • My Page
    • Receive email
      updates
    • Edit Profile
  1. NTU Theses and Dissertations Repository
  2. 工學院
  3. 土木工程學系
Please use this identifier to cite or link to this item: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/101504
Title: 基於街景影像之建築物窗牆比的自動化提取與分析
Automated Extraction and Analysis of Building Window-to-Wall Ratios Using Street View Imagery
Authors: 蔡孟璇
Meng-Syuan Tsai
Advisor: 韓仁毓
Jen-Yu Han
Keyword: 窗牆比,城市建築能源消耗影像辨識深度學習決策支援
Window-to-Wall Ratio (WWR),Urban Building Energy ConsumptionImage RecognitionDeep LearningDecision Support
Publication Year : 2025
Degree: 碩士
Abstract: 隨著氣候變遷持續加劇,實現「淨零碳排放」(Net Zero Emissions)已成為全球各國政府與產業的重要目標。建築部門約佔全球能源消耗的36%,並產生相當比例的碳排放,成為能源效率政策中最關鍵的施力點之一。其中,窗牆比(Window-to-Wall Ratio, WWR)作為建築設計中的重要參數,不僅影響自然採光與通風條件,更直接關係到建築物的熱舒適性與能源使用表現。適當的WWR設計有助於平衡日照利用與空調(Heating, Ventilation, and Air Conditioning, HVAC)需求,進而提升建築能效與室內環境品質,對於推動永續都市發展具有重要意義。然而,現行WWR資料多仰賴人工丈量或建築圖面擷取,不僅成本高、效率低,亦難以應用於都市尺度之建築能源分析。本研究以臺北市為實證場域,提出一套結合地理資訊系統(GIS)、街景影像與深度學習之自動化WWR資料擷取流程。方法首先透過Google Street View靜態API自動取得建築立面影像,並以Detectron2進行實例分割,準確辨識建築牆面與窗戶區域,再以遮罩像素統計方式計算WWR數值,並整合至空間屬性資料中。此流程可大幅減少人工介入,並具備高效率與擴展性,適用於大規模都市環境下的建築資料分析需求。本研究實際應用於建築樣式多元且街廓密集之臺北市區,驗證其可行性與自動化程度,並針對模型在標註品質、資料地區化與遮蔽物干擾等面向提出改進方向。研究成果顯示,本方法具備實用性與推廣潛力,能有效支援都市建築節能改造、碳排盤查與永續城市規劃等應用場景,為未來建築資訊擷取技術提供具體貢獻。
As climate change accelerates, achieving Net Zero Emissions has become a global priority. The building sector, responsible for about 36% of global energy use, plays a key role in carbon reduction. Among design factors, the Window-to-Wall Ratio (WWR) significantly impacts natural lighting, ventilation, and energy performance. However, conventional WWR data collection relies on manual measurement or drawings, limiting scalability.This study proposes an automated WWR extraction workflow combining Google Street View imagery, deep learning, and GIS. Using Detectron2 for instance segmentation, building façades and windows are identified from street images, and WWR is calculated via pixel-based mask statistics. The method is applied to dense urban areas in Taipei, demonstrating its feasibility and scalability.Results show that the approach effectively reduces manual effort and enables large-scale WWR analysis, supporting applications such as energy retrofitting, carbon audits, and sustainable urban planning.
URI: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/101504
DOI: 10.6342/NTU202501158
Fulltext Rights: 同意授權(限校園內公開)
metadata.dc.date.embargo-lift: 2031-01-19
Appears in Collections:土木工程學系

Files in This Item:
File SizeFormat 
ntu-114-1.pdf
  Restricted Access
25.57 MBAdobe PDFView/Open
Show full item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

社群連結
聯絡資訊
10617臺北市大安區羅斯福路四段1號
No.1 Sec.4, Roosevelt Rd., Taipei, Taiwan, R.O.C. 106
Tel: (02)33662353
Email: ntuetds@ntu.edu.tw
意見箱
相關連結
館藏目錄
國內圖書館整合查詢 MetaCat
臺大學術典藏 NTU Scholars
臺大圖書館數位典藏館
本站聲明
© NTU Library All Rights Reserved