路面缺陷包括坑洞、裂縫等，不但種類很多，嚴重程度不一，且影像形狀常與人手孔蓋類似，以致辨識效果並不理想，亟需針對各種缺陷辨識技術做進一步的強化，增加其實用性。本研究將研發影像資料自動辨識之分析架構，並分析有關影像資料之保存、管理及運用，將特定時間、特定路段的鋪面影像資料拍攝並儲存，使鋪面檢查與情況能以後製方式於室內進行，以增加管理效率，提昇鋪面檢查正確性與公信度，未來更可利用定期之路面影像資料，進行路面缺失之篩檢，以及路面成效的評估。 Surface distress including holes, cracks, etc., not only many different types but also severity of different. In this case, the image of manholes is similar as pitching and it will reduce thw accurate rate of image identification. This study will construct a architecture to analyze pavement image and find an effective way in information preservation, management and use of this image. We try to survey in a regular time, and analyze the collection data of pavement image on off-line way. We can use these results to give officers advices of pavement conservation, and judge the pavement performance. 研究期間 : 9802 ~ 9812
