自動化鋪面破壞檢測系統之研究

Automated Pavement Distress Detection System

鄧智陽(Chih-Yang Teng)

指導教授 ： 周家蓓

國立臺灣大學/工學院/土木工程學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.02455

摘要

自動化鋪面破壞檢測系統已成為世界各國管理鋪面狀況之輔助工具，配合鋪面平坦儀等資料之收集，評估道路鋪面績效，作為養護與維修參考策略之依據。我國目前以鋪面狀況指標評估鋪面損壞之情形，然而，地方道路養護管理機關因人力分配與經費不足等問題，造成鋪面管理成效不彰。因此，使用有效率的方式管理鋪面績效為未來發展之主要課題。
本研究透過實驗設計過程，針對柔性鋪面進行現地調查，收集自動化鋪面破壞檢測系統影像與人工調查資料，進行結果分析並綜合比較。鋪面破壞檢測系統因市區道路中過於頻繁之補綻與管線挖埋回填、各種干擾因素等複雜道路背景而造成結果較不理想。其經過本研究嘗試以灑水方式改善裂縫特徵，結果發現於柔性鋪面確實可增加分析軟體判斷裂縫之能力。此外，經由半自動調查和人工調查探討不同結果之差異。
本研究以半自動調查與人工調查兩種方式總花費時間進行分析，指出使用半自動總時間（平均36.49分鐘）均大於人工調查方式（平均13.03分鐘）之時間，顯示利用儀器收集影像後，進行影像瀏覽並紀錄破壞，較人工於現地觀察並紀錄破壞花費較多時間，且接近三倍。另外，使用自動化方式收集影像時，實地檢測時間則遠較人工調查少，派遣人力亦可減少。且國內交通錯綜複雜，以肉眼判斷鋪面破壞裂縫容易受到影響。因此，使用自動化方式調查可快速獲得資料且保障人員之安全。

關鍵字

鋪面自動化檢測 ； 鋪面破壞 ； 鋪面狀況指標PCI ； 統一裂縫指標UCI ； 鋪面破壞調查時間成本

並列摘要

Automated Pavement Distress Detection System (APDDS) has been a useful tool for pavement management in the world. APDDS can evaluate pavement performance and make maintenance and reconstruction（M&R）decisions based on distress and roughness data, it. In Taiwan, road authorities use pavement condition indices to evaluate pavement condition. However, insufficient labor and budget cause inefficient management works. Therefore, using efficient way to manage pavement performance is an important issue in the future.

This study investigates several flexible pavement sections and collects automated/manual pavement distress data according to experiment design. Further, the data are analyzed and compared. It was found that the data collected by APDDS are undesirable due to frequent patches and other inferences. However, this study found that the ability of detecting cracks in PicCrack could be improved by spraying water on flexible pavement. In addition, this study discussed the differences of automatically/manually-collected survey results.

This study analyzes the total time needed in semi-automated and manual survey. The results show that in all sections the total time needed in semi-automated survey (average 36.49 minutes) is longer than the total time needed in manual survey (average 13.03 minutes). This phenomenon indicates that using APDDS to collect image data takes more time because the surveyors have to go through the data and note down distresses. However, using APDDS can obviously reduce the time and labor during field survey. Moreover, manual field survey is easily affected by complex traffic condition in Taiwan. Therefore, by using APDDS, road data could be obtained quickly and the safely of surveyor could be protected.

並列關鍵字

Automated pavement distress detection ； Pavement distress ； Pavement Condition Index（PCI） ； Unified Crack Index（UCI） ； Pavement condition survey time cost

參考文獻

[1] 陳弘章，「應用慣性式平坦儀量測數據於道路整建平坦度驗收之研究」，國立臺灣大學土木工程學研究所碩士論文，民國95年10月

[10] 蕭翔駿，「整合雷射與數位影像技術於鋪面車轍量測設備之研究」，國立臺灣大學土木工程學研究所碩士論文，民國93年6月

[3] Wangxin Xiao, Xinping Yan and Xue Zhang, “Pavement Distress Image Automatic Classification Based om DENSITY-Based Neural Network”, National Basic Research Program of China, 2006, pp685-692

[4] Y. Huang and Bugao Xu, “Automatic Inspection of Pavement Cracking Distress”, Project No. 5-4975-01, Texas Department of Transportation, 2006

[5] H. D. Cheng, Jim-Rong Chen, Chris Glazier and Y. G. Hu, “Novel Approach to Pavement Cracking Detection Based on Fuzzy Set Theory”, Journal of Computing in Civil Engineering, ASCE, Vol. 13, No. 4, October 1999, pp.270-280