An Automatic Method for Detecting Side Frame Key Using Histogram of Oriented Gradient and Local Binary Pattern
Jinmin Zhang, Siming Wang, Yingke Feng, Tianxi Yang · International Journal of Control and Automation · 2017
Image processing technology plays an important role in railway fault recognition system.This paper an fault detection method based on histogram of oriented gradient(HOG) and local binary pattern(LBP) for the fault of side frame key(SFK) losing that is one of the freight car typical faults is proposed.The method contained two stages: one is the positioning of the SFK region, and another one is recognition of the fault for the SFK losing.In positioning stage, the application of the improved randomized Hough transform to extract circles on the axle and lines on the side frame get the side frame endpoint.Then, a geometric model is built to calibrate and extract the SFK region according to the geometric relationship among the axle, side frame endpoint and SFK.In the detection stage, the LBP feature histogram and HOG of ROI(region of interest) were statistically respectively, using principal component analysis(PCA) for dimensionality reduction, and the two histogram to constitute a joint histogram.Finally, using support vector machine(SVM) classification, fault detection.Experimental results show that this method the average fault detection rate is 96.15%, and the method has good applicability.