Subway door-punching behavior detection method and system based on neural network

A technology of neural network and detection method, which is applied in the field of detection method and detection system of subway door-slamming behavior based on neural network, and can solve the problem of low reliability of passengers' door-slamming behavior detection

CN111985453AInactive Publication Date: 2020-11-24赵华
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2020-11-24
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to a subway door-punching behavior detection method and system based on a neural network. The method comprises the steps of obtaining a training set, and building a neural network model; training the neural network model by taking the images marked with key points in the training set as input and taking the corresponding key point skeleton as output, and taking the trained neural network model as a posture recognition neural network model; when the subway door is in a nearly closed state, acquiring an image of a set area around the subway door, and inputting the image into the posture recognition neural network model to obtain a key point skeleton of a target in the image; and judging whether the target moves towards the subway door or not according to the key point skeleton of the target in the image, if so, obtaining the moving speed of the target, and judging whether the behavior of rushing to the subway door exists or not according to the moving speed. According to the technical scheme provided by the invention, the problem of low reliability of passenger door punching behavior detection in the prior art can be solved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of subway safety detection, and in particular relates to a neural network-based detection method and detection system for door-slamming behavior of subways. Background technique

[0002] The subway operates on fully enclosed lines. The lines located in the central urban area are basically set up in underground tunnels, and the lines outside the central urban area are generally set up on viaducts or on the ground. It is a special right-of-way covering various underground and above-ground areas in urban areas. There is a high-density, high-capacity urban rail transit system. Due to the characteristics of fast speed, large transportation volume, stable time point, and economic benefits, the subway is very popular in major cities.

[0003] During the morning rush hour in the city, passengers rush to the door on almost every subway train. The station attendants usually take measures such as blowing whistles and ...

Examples

Embodiment Construction

[0042] The object of the present invention is to provide a method and system for detecting door-slamming behavior of subways based on neural network, so as to solve the problem of low reliability in detection of passenger's door-slamming behavior in the prior art.

[0043] Method example:

[0044] This embodiment provides a method for detecting the behavior of rushing the subway door based on a neural network, the process of which is as follows figure 1 shown, including the following steps:

[0045] Step 1: Obtain a gesture recognition neural network model, which is used to obtain the key point skeleton of the target in the image.

[0046] Step 2: When the subway door is close to closing, continuously acquire multiple images within the set area around the subway door, input them into the gesture recognition neural network model, and obtain the key point skeleton of the target in each image. The target in the image mentioned in this embodiment is the passenger in the image. ...