Metro Security Inspection Method Based on Spatial Feature Transformation Network

In order to ensure the safety of urban rail transit, X-ray baggage inspection systems have been widely used. However, at the current, baggage inspection mainly relies on manual judgment, which has the problems of high labor cost, low efficiency and lack of objectivity. This paper aims to realize intelligent and automated baggage inspection and studies and designs a dangerous goods detection method based on spatial feature transformation network. This method adopts target detection technology based on Anchor mechanism, and transforms high-level semantic features spatially spatial transformation network layer. The feature pyramid fusion approach is employed to integrate the transformed high-level semantic features with the low-level detailed features, resulting in the generation of multi-scale features that are rich in semantic information. This innovative method enables accurate prediction of the categories and precise localization of various hazardous items within X-ray images, thereby significantly enhancing the detection accuracy.It proves the feasibility of using intelligence algorithms for automatic security inspection, and provides strong support for the implementation of automatic monitoring of urban rail transit baggage inspection.