Internal Detection of Ground-Penetrating Radar Images Using YOLOX-s with Modified Backbone

Geological radar is an important method used for detecting internal defects in tunnels. Automatic interpretation techniques can effectively reduce the subjectivity of manual identification, improve recognition accuracy, and increase detection efficiency. This paper proposes automatic approach geological images (GPR) based on YOLOX-s, aimed at accurately steel arches any direction. The utilizes YOLOX-s neural network improves backbone with Swin Transformer to enhance capability small targets images. To address irregular voids commonly observed images, CBAM attention mechanism incorporated accuracy annotations. We construct a dataset using field data that includes different sizes orientations, representing “voids” “steel arches”. Our model tackles challenges traditional GPR image enhances efficiency detection. In comparative experiments, our improved achieves 92% 94% arches, as evaluated constructed dataset. Compared average precision by 6.51%. These results indicate superiority interpretation.