作者: Yunus Santur , Mehmet Karakose , Erhan Akin
DOI: 10.1109/INDIN.2016.7819184
关键词: Condition monitoring 、 Test phase 、 Integral method 、 Simulation 、 Random forest 、 Feature extraction 、 Fuzzy logic 、 Real-time computing 、 Work (physics) 、 Engineering 、 Real systems
摘要: Today, railway systems are one of the most preferred means transportation worldwide. Some faults may occur on tracks due to several reasons and such cause accidents. For this reason, should be periodically inspected. In study, a study was performed which aimed locate possible as cracks, holes wear rail surfaces fast with high accuracy levels in accordance real systems. The is made two stages distance values read laser cameras constitute input data system. Rail profiles marked working out-of-order presented system random forest created for learning after pretreatment data, four different feature extraction methods obtained attribute. test phase, second attribute via methods, classified produced diagnosis. General rates increased fuzzy integral method can work time railways physical tester proposed.