The accurate positioning of bridge components is helpful to the health monitoring and maintenance management of bridge structures. However, the traditional bridge inspection often requires a lot of manpower and time, and the detection efficiency and accuracy are also very low. To address this issue, this paper proposes a bridge component object detection method based on improved YOLOV7. Firstly, this paper created the Bridge Component Dataset (BC-Dataset), consisting of 1200 images, including three types of structures: bridges, piers, and girders. Secondly, in order to enhance detection accuracy in complex scenes, this paper incorporated the dynamic sparse attention (the biformer attention) module and Wise-IoU into the YOLOV7 network. Finally, experiments were conducted on the BC-Dataset, achieving a model accuracy of 91.19%. The experimental results demonstrate that our method outperforms traditional methods in terms of precision and accuracy in bridge component object detection tasks. This research is of great significance for the safety and monitoring of bridge structures, providing strong support for research and applications in related fields.