Signal planar graph signal machine identification method and system based on deep learning

Abstract

The invention provides a signal plane graph signal machine identification method and system based on deep learning. The method comprises the following steps: acquiring a signal plane graph to be identified; identifying a signal machine in the signal plane graph by adopting a signal machine identification network; and the signal machine identification network is a neural network based on YOLOv5-s. According to the signal planar graph signal machine identification method and system based on deep learning, a Signal-Net identification network suitable for a signal planar graph is provided on the basis of a YOLOv5-s network, and a signal machine target in a high-resolution drawing can be accurately detected. The model parameter quantity is moderate while the identification precision of the network on the signal machine is ensured, and the requirement of real-time operation on the edge computing equipment is met.

Description

technical field [0001] The invention belongs to the field of rail transit, and in particular relates to a method and system for identifying signal planes based on deep learning. Background technique [0002] For the signal field of rail transit management, in the process of setting up the drawing plan, it mainly includes drawing approach signals, warning signs, and station signals, etc., and they must meet the relevant requirements. In view of the many types of signal machines in the station, even the same type of signal machine needs to be composed of signal machine mechanisms in different ways according to the actual situation of the on-site lines. All kinds of signals are time-consuming and labor-intensive work. [0003] As one of the important algorithms for target detection, YOLOv3 (YOLO, You Only Look Once), because of its high precision and fast reasoning speed, enables engineers and technicians to see the dawn of computer vision deployed in the industrial field. As ...

AI Technical Summary

Problems solved by technology

[0012] In view of the above-mentioned problems, the present invention aims at the problem that the current deep neural network is difficult to detect very small targets in large resolution or super-large resolution, adopts deep learning target detection algorithm combined with traditional image processing methods, and uses the signal machine in the signal plane as the Recognize the target, design a network for signal machine identification in the signal plane, that is, the Signal-Net network, to identify signal machines to improve the efficiency and accuracy of manual identification

Images