Abstract: To effectively address the issues of low efficiency and insufficient accuracy in manual inspection and acceptance during EMU bogie maintenance, this paper develops an intelligent handover inspection system for bogie maintenance that integrates emerging technologies such as intelligent multi-robot control, image recognition and analysis, and deep learning. The system achieves automatic and precise positioning of bogies through push positioning robots, completes comprehensive image collection of the externally visible areas of bogies by top and bottom image acquisition robots, and conducts in-depth recognition and analysis via the intelligent detection subsystem. Integrating key technologies including YOLOv5-based object detection algorithm, semantic segmentation-based lockwire detection algorithm, and deep learning-based multi-type defect detection model, the system enables intelligent inspection of four typical types of bogies used in CRH380B and CR400BF EMUs. Through continuous iterative optimization, the test results show that the system achieves a defect detection recall rate of over 99% and an accuracy rate of 95% for the key externally visible components of bogies. Compared with traditional manual inspection, it can improve inspection efficiency by 48.5% per bogie. This system can effectively replace manual inspection in the final assembly acceptance process of bogies.