Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Railway tunnel crack detection method based on improved residual network

A detection method and tunnel technology, applied in neural learning methods, biological neural network models, image data processing, etc., can solve problems such as loss of details, missed detection of small cracks, failure to find image features of railway tunnels, etc., to achieve good classification and increase The effect of accuracy

Pending Publication Date: 2021-04-30
BEIJING UNION UNIVERSITY
View PDF8 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that it cannot find the image features of small cracks in the railway tunnel under the complex background, and it cannot restore the detailed structure information of the image after the pooling layer. There are problems of edge smoothing and loss of details, which may easily cause small cracks to be missed. The phenomenon

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
  • Railway tunnel crack detection method based on improved residual network
  • Railway tunnel crack detection method based on improved residual network
  • Railway tunnel crack detection method based on improved residual network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0049] Such as figure 1 As described above, step 100 is executed to collect images of railway tunnel cracks and obtain a railway tunnel crack detection data set.

[0050] Execute step 110, perform enhancement processing on the images in the railway tunnel crack detection data set and randomly divide the images into a training set and a test set. The enhancement processing includes performing at least one of gray scale processing, Gaussian filtering and image normalization on the image.

[0051] Execute step 120 to improve the residual network structure, use the images in the training set to train the improved residual network, continuously optimize the residual network structure through training, and use the test set images to test the accuracy of the residual network after the training is completed . The improved residual network structure is that the pyramid hole convolution module is integrated into the bottom of ResNet, the number of convolution kernels of each convoluti...

Embodiment 2

[0072] Aiming at the characteristics of the railway tunnel crack image, the present invention proposes a railway tunnel crack detection method based on the improved residual network (Pyramid Dilated Convolution Residual Network, PDC-ResNet), the method improves the ResNet residual network structure, and the main improvement aspects are : Combine dilated convolution blocks with different expansion rates with traditional convolutional blocks to form a pyramidal dilated convolution module; add a pyramid dilated convolution module to improve the underlying receptive field of the ResNet network and improve classification accuracy; use metrics-based The learned combination loss function is used to distinguish similar differences between different classes, reduce the missed detection rate and false detection rate of cracks, and thus better realize the detection of small cracks in complex backgrounds.

[0073] To achieve the above object, the present invention is achieved through the f...

Embodiment 3

[0107] The present invention provides a railway tunnel crack detection method based on the improved residual network. Compared with the existing technology, it has the following obvious advantages and beneficial effects:

[0108] (1) Add pyramidal hole convolution modules with different expansion rates at the bottom of ResNet to ensure that the resolution of the feature map is not reduced, and the receptive field of the convolution kernel can be expanded, which can be well used for the extraction of multi-scale features of tunnel crack images , to increase the classification accuracy.

[0109] (2) By designing a combined loss function based on metric learning, the model can increase the distance between different classes as much as possible during training to better classify small cracks in complex backgrounds.

[0110] (3) Compared with the ResNet basic network, the present invention can improve the accuracy of railway tunnel crack identification, and can effectively and time...

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 provides a railway tunnel crack detection method based on an improved residual network. The method comprises the steps: collecting a railway tunnel crack image, and obtaining a railway tunnel crack detection data set; the method also comprises the following steps: carrying out enhancement processing on images in the railway tunnel crack detection data set and randomly dividing the images into a training set and a test set; improving the structure of the residual network, and training the improved residual network by using the images in the training set; and carrying out crack classification detection on the railway tunnel image by using the trained residual network structure. According to the method, a ResNet residual network structure is improved to form a pyramid cavity convolution module; a pyramid cavity convolution module is added to improve the underlying receptive field of the ResNet network and improve the classification accuracy; a combined loss function based on metric learning is adopted to distinguish similar differences between different classes, the omission ratio and the false drop rate of cracks are reduced, and therefore detection of small cracks under a complex background is better achieved.

Description

technical field [0001] The invention relates to the technical field of railway intelligent monitoring and machine vision, in particular to a railway tunnel crack detection method based on an improved residual network. Background technique [0002] During the operation period of the railway tunnel, there are different degrees of diseases due to various reasons. The cracks in the lining are the most common one. The cracks will affect the stability of the tunnel, which is a great hidden danger to the safe operation of the railway tunnel. Therefore, timely and effective identification and treatment of tunnel cracks is a very important task. At present, the detection of cracks in railway tunnels mostly uses the method of manual inspection, which is difficult to meet the requirements of the rapid development of railway safety detection. [0003] At present, the use of computer vision and digital image processing to detect cracks has attracted more and more attention. Commonly use...

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): G06T7/00G06N3/04G06N3/08
CPCG06T7/0008G06N3/08G06T2207/20081G06T2207/30108G06T2207/10004G06N3/045
Inventor 饶志强常惠李益晨李子仡赵玉林
Owner BEIJING UNION UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com