Upper pull rod fault detection method based on deep learning

A technology of fault detection and deep learning, applied in the direction of railway vehicle shape measuring devices, biological neural network models, instruments, etc., can solve the problems of missed detection detection efficiency, error-prone upper pull rods, etc., to improve detection accuracy, improve quality and The effect of detection efficiency and labor cost saving

Active Publication Date: 2021-03-02
HARBIN KEJIA GENERAL MECHANICAL & ELECTRICAL CO LTD
View PDF10 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problems of prone to errors, missed detection and low detection efficiency when using the existing method to detect the fault of the upper rod, and propose a method for detecting the fault of the upper rod based on deep learning

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Upper pull rod fault detection method based on deep learning
  • Upper pull rod fault detection method based on deep learning
  • Upper pull rod fault detection method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0028] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS 1. This implementation will be described with reference to FIG. 1 . A method for detecting a pull-up rod fault based on deep learning in this embodiment, the method is specifically implemented through the following steps:

[0029] Step 1, collect the train image to be detected, and obtain the image of the region of interest from the collected train image;

[0030] Set up high-definition equipment at the bottom of the railway freight car tracks to take pictures of trains passing at high speed and obtain high-definition images of the bottom of the car body. Using line scanning, seamless stitching of images can be realized, and a two-dimensional image with a large field of view and high precision can be generated. According to the wheelbase information, bogie type and vehicle type of the EMU, the upper tie rod is roughly positioned, and the local area image containing the upper tie rod parts is cut out from the captured pi...

specific Embodiment approach 2

[0037] Specific embodiment two: the difference between this embodiment and specific embodiment one is: in step six, the FasterRCNN model that contains the image to be recognized that is extracted in step five and includes the upper tie rod is input into the trained FasterRCNN model, if the FasterRCNN model detects the upper tie rod joint and For the position of the flat iron, determine whether the upper tie rod has fallen off according to the connection status of the upper tie rod joint and the flat iron;

[0038] If there is a failure of the upper tie rod falling off, a message will be generated and uploaded to the alarm platform; if there is no failure of the upper tie rod falling off, the image to be recognized (that is, the image after enhanced processing) will be input into the trained Unet semantic segmentation model, and the trained Unet semantic segmentation model will be used. The Unet semantic segmentation model segmented the upper tie rod joint and the upper tie rod ...

specific Embodiment approach 3

[0039] Embodiment 3: The difference between this embodiment and Embodiment 1 is that in step 2, the image of the region of interest is enhanced, and the image of the region of interest after the enhancement is obtained. The specific process is as follows:

[0040]

[0041]

[0042]

[0043] Among them, v(x, y) represents the gray value of the pixel point (x, y) in the image of the region of interest, and I 2 (x, y) represents the gray value of the pixel point (x, y) of the image of the region of interest after nonlinear transformation, Represents the average gray value of all pixels in the image of the region of interest, m(x, y) and kv(x, y)) are intermediate variables, and a is the adjustment coefficient;

[0044] The smaller a, the larger the gray value of pixels with small gray values ​​after nonlinear transformation, but at the same time it is easy to lose texture detail information; the larger a is, although the texture information is retained, but the pixels w...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an upper pull rod fault detection method based on deep learning, and belongs to the technical field of upper pull rod fault detection. According to the method, the problems ofhigh possibility of errors and missing detection and low detection efficiency when the existing method is used for detecting the faults of the upper pull rod are solved. The method comprises the following steps: firstly, acquiring a train image to be detected, and acquiring a region-of-interest image from the acquired image; carrying out enhancement processing on the region-of-interest image to obtain an enhanced region-of-interest image; after the image is segmented, outputting the brake cylinder part positions in the segmented sub-images through an improved SSD model; extracting a to-be-identified image containing the upper pull rod according to the position of the brake cylinder; and finally, carrying out fault identification on the pull-up rod in the image to be identified by adoptinga FastRCNN model and a Unet model. The method can be applied to upper pull rod fault detection.

Description

technical field [0001] The invention belongs to the technical field of failure detection of upper tie rods, and in particular relates to a method for detecting failures of upper tie rods based on deep learning. Background technique [0002] The upper tie rod is an important part of the braking system of the railway vehicle, which plays the role of transmitting the braking force when the train is braking. When the vehicle brakes, the braking force output by the brake cylinder is transmitted to the upper tie rod through the brake lever. The upper tie rod is connected with the bogie moving lever through the round pin. brake. If the upper link breaks during the braking process, the braking force cannot be transmitted to the brake shoe, the vehicle loses its braking ability, and cannot stop, resulting in a driving safety accident. [0003] At present, the vehicle inspection method is mainly adopted by manually looking at pictures one by one. This method is affected by factors s...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/32G06K9/34G06N3/04B61K9/00
CPCB61K9/00G06V10/25G06V10/267G06N3/045
Inventor 马元通
Owner HARBIN KEJIA GENERAL MECHANICAL & ELECTRICAL CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap