國內自民國88年引進ARRN檢測技術後，藉由各方學者與專家之努力，國內發展出許多的鋪面維護管理系統與檢測設備，而為使鋪面維護管理系統之資訊能不斷快速更新，定期定時的提供大量的鋪面資訊對於系統是不可或缺的，而要如何使用檢測儀器完成路面影像掃瞄、拍攝、量測等工作，並使用軟體自行判讀計算指標，偵測鋪面破壞予以定位，在國內管理單位並無實際執行成功案例。 本研究成功研發出智慧型鋪面檢測設備，該設備能進行鋪面縱向剖面、鋪面影像及道路附屬設施等影像進行紀錄，並將影像結果轉換為定量指標提供鋪面維護管理系統作為分析之依據，並將影像資料之保存、管理及運用作業標準化，使鋪面檢查與情況審視工作得以快速執行，提昇鋪面檢查正確性與公信度。 平整度量測系統依據ASTM E950與AASHTO PP49等相關規定將水準儀、ARRB與智慧型鋪面檢測設備進行量測結果比較，並進行檢定判斷檢測車之IRI值與ARRB無顯著差異，建議未來驗證次數以12次做為驗證之標準次數；影像辨識軟體能以自動化方式進行鋪面指標計算(Unified Condition Index、Universal Condition Index)，偵測定位破壞位置(坑洞、鱷魚狀裂縫、橫向裂縫、縱向裂縫、補綻等)，並利用資料採礦技術建立自動指標與鋪面狀況指標(Pavement Condition Index ,PCI)之分類準則，未來可利用自動指標計算結果判斷鋪面狀況；道路影像實錄系統除能將附屬設施影像進行定位，本研究尚利用定框技術建立影像自動匯入功能，能將匯入影像自動進行分類，藉以統計數量與種類。 本研究使用智慧型鋪面檢測設備進行桃園縣鄉道資料蒐集，利用灰預測與回歸模式建構桃園縣劣化模式，再使用熵值法取得縱向裂縫指標、橫向裂縫指標、鱷魚狀裂縫指標、坑洞與補綻指標之間權重，建構鄉道鋪面現況指標CPPI，希望藉此能客觀評估鋪面服務力現況；此外，依據平整度普查資料與桃園縣鄉道養護能量訂定養護平整度門檻值為7.88 m/km，並使用鋪面狀況推估估計桃園縣鄉道使用年限為5年。 整合上述成果，將各分析模式結合路面各項資料與路網圖形化資料，發展桃園縣鋪面養護管理地理資訊系統，能使用該系統執行鋪面管理的理論、方法和實務經驗，並可提供實務單位確實可行的管理策略與分析工具。 Since 1999, there were a lot of studies about pavement management system, but unfortunately we did not use it in real maintainance works. The purpose of this study is that we want to integrate three kinds of important technique into intelligent survey vechiel, and try to think a standard operation for pavement survey. We did a lot of tests and verified about the pavement roughness measurement system in accordance with ASTM E950 and AASHTO PP49. We find that all results of test is conforming the standard of ASTM and AASHTO. Roughness is a very important performance indicator in pavement maintenance management systems. The international rough index of indicators was used as the core of roughness measuring of degree gradually in recent years domestically, but to predict a smooth degree of indicators the method is set up differently as there is a different prediction indicator because the theoretical foundation is different.We predicts the deteriorationof IRI by different ways, including grey forecast, multiple regression. We also try to use data minning technique to find a rule between unified condition index and pavement condition index, and the accuracy of the J48 classification tree is 80%. This study uses the pavement data of country road in Taoyuan County to build “Country Pavement Performance Index” by entropy. Now, it is a simple way to know the profiler of pavement and detect pavement distress at the same time. This intellegent system can use vediologging module to locate the road image and build a function to let road image import into pavement management system automatically. We can supply the latest information of road to pavement management system.
