Research and Implementation of Deep Learning-based Interlocking Upper Computer Interface Recognition Methods

Aiming at the problem that the interlocking station data required by the automatic computer interlocking test software relies on manual entry by hand and has a low accuracy rate, this paper proposes a deep learning-based computer interlocking software uplink interface recognition method, which analyzes the signaling equipment and text display characteristics by intelligently recognizing the interlocking uplink interface that is verified to be free of errors. Using the improved YOLOv10s network framework and deep learning related algorithms, signaling equipment and text information such as signaling machines and turnouts in the interlocking station diagram are accurately identified and the recognition task is completed efficiently and accurately, which greatly improves the testing efficiency.