Modification of Deeplab V3+ for Detecting Cracks on Steel Rail Surfaces

This study proposes an improved deep learning model for detecting surface cracks on railway steel tracks, addressing the urgent safety needs in the growth of railway construction. The improved Deeplab V3+ model architecture can improve feature extraction efficiency and reduce computational costs. Based on experimental results, the model outperforms others in pixel accuracy and processing speed. This study provides an effective approach for improving railway crack detection, and future work will focus on further optimizing the model for practical applications in railway inspections.