Vehicle and Parking Space Detection Based on Improved YOLO Network Model

YOLO has a fast detection speed and is suitable for object detection in real-time environment. This paper is based on YOLO v3 network and applied to parking spaces and vehicle detection in parking lots. Based on YOLO v3, this paper adds a residual structure to extract deep vehicle parking space features, and uses four different scale feature maps for object detection, so that deep networks can extract more fine-grained features. Experiment results show that this method can improve the detection accuracy of vehicle and parking space, while reducing the missed detection rate.