Rail Surface Defect Detection and Location Based on Image Processing

Abstract

In recent years, railways have become an important means of transportation for the country, bringing great convenience to the masses and promoting the economic development if the country. These rails play a crucial part in the train operating. The detection of rail surface defects is vital for high-speed rail maintenance and management. The CNN-based computer vision approach has been proved to be a strong detection tool widely used in various industrial scenarios. However, the CNN-based detection models are diverse from each other in performance, and most of them require sufficient training samples to achieve high detection performance. Selecting an appropriate model and tuning it with insufficient annotated rail defect images is time-consuming and tedious. With an aim of conquering the challenge, stimulated by ensemble learning which employs the multiple learning algorithms so as to obtain greater predictive performance, we create one ensemble framework for detecting the industrialized rail defect. As well as image augmentation operation, feature augmentation operation is adopted to deeply get the model more diverse at random. A shared feature pyramid network is adopted to reduce model parameters as well as computation cost. Experimental results substantiate that the approach out-performs single detecting architecture in our specified rail defect task. On the collected dataset with 8 defect classes, our algorithm achieves 7.4% higher mAP.5 compared with YOLOv5 and 2.8% higher mAP.5 compared with Faster R-CNN.