Railway Fastener Positioning Method Based on Improved Census Transform

In view of the fact that the current positioning methods of railway fasteners are easily affected by illumination intensity, bright spots, and shadows, a positioning method with relative grayscale invariance is proposed. The median filter is used to remove the noise in order to reduce the adverse effects on the subsequent processing results, and the baffle seat edge features are enhanced by improved Census transform. The mean-shift clustering algorithm is used to classify the edges to weaken the interference by short lines. Finally, the Hough transform is used to quickly extract the linear feature of the baffle seat edge and achieve the exact position of the fastener with the prior knowledge. Experimental results show that the proposed method can accurately locate and have good adaptability under different illumination conditions, and the position accuracy is increased by 4.3% and 8%, respectively, in sunny and rainy days.