Research On High Speed Rail Fastener Detection Algorithm Based On Component Model

With the rapid development of China’s high-speed rail industry,railway track detection has become increasingly important as an important basis for ensuring the safe and stable operation of trains.In recent years,image processing technology has developed rapidly and is gradually being used for automated detection of railway tracks.The fastener is a key part of fixing the rail to the sleeper.If the loss or breakage will cause serious safety accidents during the train operation,the research on the detection algorithm of the fastener is of great significance.The deformable component model is a component-based target detection model,which is widely used in the field of target recognition.This paper tests the high-speed rail fasteners based on the deformable component model.The main contents are as follows:(1)In order to realize the accurate,fast and automatic detection of high-speed iron fasteners,a fastener detection algorithm combining feature similarity and component model is proposed.Firstly,the high-speed rail fastener is trained by using the hidden variable SVM to obtain the fastener component model and the weight of each component.Secondly,the reasonable model pyramid layer number is determined according to the fastener characteristics,and the fastener component model is masked to remove the part.The background area;then,the non-mask area HOG feature vector is calculated,and the fastener is accurately positioned according to the feature vector and the component weight;finally,since there are multiple fasteners in each image,the distance information of the adjacent fasteners is calculated.And the feature similarity as a quick judgment criterion for the existence of the defective fastener,determining whether there is a defective fastener in the image.Theoretical analysis and experimental results show that the proposed algorithm has low false detection rate and missed detection rate,high detection accuracy,fast detection speed,good robustness to external illumination changes,and can meet the needs of high-speed fastener automatic detection application.(2)Aiming at the image of defective fasteners,in order to accurately identify the category of defective fasteners,a fastener detection algorithm based on multi-threshold component model is proposed.First,using the fastener component model after the mask operation to calculate multiple thresholds,the threshold interval of the track plate,the lost fastener,the suspicious fastener,and the normal fastener isobtained;then,for the image with the defective fastener,the threshold interval is used as the threshold interval The fastener detection evaluation standard detects the state of the fastener in the image;finally,for the detected suspicious fastener,the feature similarity with the normal fastener is calculated to determine the state(missing,breaking,normal)of the suspicious fastener.Theoretical analysis and experimental results show that the proposed algorithm can accurately identify the type of defective fasteners.Compared with other mainstream fastener detection algorithms,the false detection rate and missed detection rate are low,and the illumination is robust.