Back to Search
Start Over
DEVELOPMENT OF SEMI-AUTOMATIC ASPHALT PAVEMENT CRACK DETECTION SYSTEM USING IMAGE PROCESSING AND MACHINE LEARNING APPROACH
- Source :
- Journal of Japan Society of Civil Engineers, Ser. E1 (Pavement Engineering). 71:I_31-I_38
- Publication Year :
- 2015
- Publisher :
- Japan Society of Civil Engineers, 2015.
-
Abstract
- For appropriate road maintenance and management of the road, detection and evaluation of the asphalt distress is very important. In Japan, crack ratio, amount of rutting, and flatness are investigated to evaluate the damage. This research focuses on the crack ratio. Because there is no established method to detect cracks automatically, drawing and counting cracks manually from the photo of the road surface are required though enormous effort and time are needed. To solve the problem, there exist several automatic crack detection methods, but all of them suffer from the error including false detection. Therefore this research has developed the semi-automatic method which enables us to modify the automatic crack detection result manually. The developed method is tested using the photos of asphalt pavements, and it is found that the method can detect cracks with very high accuracy. ひび割れ率はアスファルト舗装の損傷を定量的に評価する指標の1つである.その算出にあたっては,路面のひび割れをスケッチした後に計算を行う必要があるが,手作業となるため膨大な労力が必要となっており,さらにはひび割れ開口幅などの重要な情報を得ていないという問題もある.そこで様々な自動化手法が考案されているが,どのような精度のよい手法であっても一定の誤検出や未検出,人間との判定の相違は避けられない.そこで本研究では自動検出手法により得られた結果を人力で微調整することの出来る半自動検出手法を構築した.この手法は画素単位でひび割れを判定できるため,ひび割れ開口幅や面積などの計算が容易であるという利点もある.そして本手法を舗装撮影画像に適用した実験により,ひび割れ検出における本手法の有用性を確認した.
Details
- ISSN :
- 21856559
- Volume :
- 71
- Database :
- OpenAIRE
- Journal :
- Journal of Japan Society of Civil Engineers, Ser. E1 (Pavement Engineering)
- Accession number :
- edsair.doi...........100d2ea4c5f215bd89693d2ce71b6a3b
- Full Text :
- https://doi.org/10.2208/jscejpe.71.i_31