Abstract: The subway driver can not complete the foreign objects detection of the gap between the door of the train and the platform screen door (PSD) before driving. This may cause operational safety problems. This paper proposed an automatic detection and early warning system of foreign objects in subway curve platform. A micro camera was installed above the PSD to photograph the gap between the PSD and the door of the train. The advanced image processing algorithm was used to detect the door of the image and determine the area of the door in the image (i.e. the region of interest). After the passengers got on and got off, the foreign object detection was carried out in the region of interest. In the presence of foreign objects, it was issued a warning to remind the driver delayed driving. The experiment proves that the system can accurately detect the presence of foreign objects in the gap, which can assist the train driver to detect the foreign objects in the gap of subway curve platform before driving.