Floor damage fault image identification method

A technology for image recognition and flooring, applied in character and pattern recognition, instruments, biological neural network models, etc., can solve problems such as low accuracy, low efficiency, and high cost, and achieve the goals of improving efficiency, reducing safety hazards, and saving labor costs Effect

Inactive Publication Date: 2020-04-28
HARBIN KEJIA GENERAL MECHANICAL & ELECTRICAL CO LTD
View PDF6 Cites 11 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 high cost, low efficiency, and low accuracy in the existing manual detection of floor damage faults of railway wagons, and propose a floor damage fault image recognition method

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
  • Floor damage fault image identification method
  • Floor damage fault image identification method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0025] Specific implementation mode one: the specific process of the floor damage fault image recognition method in this implementation mode is as follows:

[0026] Step 1. Collect samples and create data sets;

[0027] Step 2: Perform preprocessing on the image to be recognized collected by the device, and obtain the image to be recognized after preprocessing; the specific process is:

[0028] Because in the actual detection process, due to the harsh working environment of the truck floor, the collected pictures will produce various noises. When used as training samples, the noise in the pictures will be extracted as feature values, which will seriously interfere with the recognition results. In view of this situation, Gaussian filtering and gray level equalization operations are performed on the collected images to be recognized to remove noise;

[0029] Step 3: Establish a VGG model, use the training set to train the VGG model, and obtain a pre-trained VGG model; the speci...

specific Embodiment approach 2

[0040] Specific embodiment two: the difference between this embodiment and specific embodiment one is that the samples are collected in the step one, and the data set is established; the specific process is:

[0041] Step 11. Use a high-resolution line-scan camera to collect a clear grayscale image of the railway wagon floor; the principle of collecting the image is as follows:

[0042] (1) Collect the floor images of railway wagons under various conditions such as rainwater, ice and snow, chalk graffiti, load leakage, mud stains, oil stains, black paint, and dust;

[0043] (2) Collect floor images of railway wagons at different stations, different equipment, and at different times (different degrees of sunlight interference);

[0044] (3) According to the material of the railway wagon floor, images of wagon floors such as steel floors and wooden floors are collected respectively;

[0045] Step 12. Sample amplification: Due to the limited sample data actually collected, in or...

specific Embodiment approach 3

[0047] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is that the VGG model is established in the step three one; the specific process is:

[0048] VGG model structure such as figure 2 shown;

[0049] VGG model includes input layer, 64-channel conv2 convolutional layer 1, 64-channel conv2 convolutional layer 2, pool_1 pooling layer, 128-channel conv3 convolutional layer 1, 128-channel conv3 convolutional layer 2, pool_2 pooling layer, 256 channels conv4 convolutional layer 1, 256-channel conv4 convolutional layer 2, 256-channel conv4 convolutional layer 3, 256-channel conv4 convolutional layer 4, pool_3 pooling layer, 512-channel conv5 convolutional layer 1, 512-channel conv5 convolutional layer 2 , 512-channel conv5 convolutional layer 3, 512-channel conv5 convolutional layer 4, pool_4 pooling layer, 512-channel conv5 convolutional layer 5, 512-channel conv5 convolutional layer 6, 512-channel conv5 convolutional layer 7,...

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 relates to a fault image recognition method, in particular to a floor damage fault image recognition method. The objective of the invention is to solve the problems of high cost, low efficiency and low accuracy of an existing manual detection mode for the damage fault of the railway freight car floor. The method comprises the following steps: 1, collecting samples, and establishing adata set; 2, preprocessing a to-be-identified image acquired by the equipment to obtain a preprocessed to-be-identified image; the method comprises the following specific steps: performing Gaussian filtering and gray scale equalization operation on a collected image to be identified, and removing noise; 3, establishing a VGG model, and training the VGG model by adopting the training set to obtaina pre-trained VGG model; 4, inputting the preprocessed to-be-recognized image into a pre-trained VGG model, carrying out the recognition of the damage fault of the railway freight car floor, and obtaining a detection result; and 5, judging whether a detection result conforms to floor fault characteristics or not. The method is applied to the field of fault image recognition.

Description

technical field [0001] The invention relates to a fault image recognition method. Background technique [0002] The floor damage of railway wagons is a kind of failure that endangers the safety of trains. The traditional detection method is to manually check the image data collected by the side of the rail. As the inspectors are prone to fatigue and omissions during the work process, human misjudgments and missed judgments are caused, thereby affecting driving safety. At the same time, a large amount of manpower and material resources are invested in the inspection of pictures, which causes a great waste of resources for users. [0003] Therefore, in order to promote the rapid development of railway transportation automation and solve the problems of high cost, low efficiency, and low accuracy in manual detection, it is of great practical significance to realize the automation of truck fault detection. Contents of the invention [0004] The object of the present inventio...

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): G06K9/00G06K9/62G06N3/04
CPCG06V20/10G06V20/588G06N3/045G06F18/214
Inventor 生田野
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