With the rapid expansion of high-speed railway network, the real-time monitoring of trackside equipment becomes particularly important. To detect trackside equipment information more accurately, a YOLO-R algorithm grounded on the improved You Only Look Once v3 (YOLOv3) algorithm is proposed, and the trackside equipment identification and detection model is constructed. By introducing feature pyramid network and adaptive Bessel curve network, the new model can effectively identify and locate different types of trackside equipment such as switch machine, derailer, and shaft counter. The experiment findings denote that the new model is superior to the existing technology in all aspects of on-orbit equipment recognition and detection, the computer resource occupancy rate is only 22%, the image recognition accuracy rate is more than 98%, and the processing speed is up to 200 images/second. This research not only raises the automation level of trackside equipment monitoring, but also provides a powerful technology for railway safety operation.