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

Tunnel lining crack image segmentation method

An image and lining technology, applied in the field of image recognition, can solve problems such as the inability to accurately obtain crack attribute characteristics, achieve the effects of improving recognition accuracy and recognition efficiency, reducing false alarm rate, and avoiding the influence of subjective factors

Pending Publication Date: 2019-07-16
新而锐电子科技(上海)有限公司
View PDF7 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, such recognition accuracy cannot accurately obtain the attribute characteristics of cracks, such as the direction, length, and width of cracks, etc.

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
  • Tunnel lining crack image segmentation method
  • Tunnel lining crack image segmentation method
  • Tunnel lining crack image segmentation method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0049] Such as figure 1As shown, Embodiment 1 of the present invention provides a method for segmenting tunnel lining crack images, and the method includes the following process steps:

[0050] Step S110: Divide the directly collected original crack image into multiple image blocks, and select the image block with cracks and the corresponding binary image as the training sample set;

[0051] Step S120: De-mean and normalize the image blocks with cracks in the training sample set;

[0052] Step S130: using the convolutional neural network and combining the loss function to train the training sample set after de-mean and normalization processing to obtain a crack segmentation model;

[0053] Step S140: Input the fracture image to be segmented into the fracture segmentation model after demeaning and normalizing, and output the segmentation result image.

[0054] In Embodiment 1 of the present invention, the step S110 specifically includes:

[0055] Each original crack image di...

Embodiment 2

[0077] Such as Figure 5 As shown, the training process and testing process of a deep neural network model for crack region segmentation in tunnel lining images provided by Embodiment 2 of the present invention.

[0078] In the second embodiment of the present invention, a large amount of detailed information needs to be extracted to finely segment the slender crack region. Therefore, the neural network of the present invention incorporates the low-dimensional features (edges, gradients, etc.) extracted during the training process. And high-dimensional features (semantic features of cracks, etc.), through training, the final high-precision segmentation effect is obtained.

[0079] For the directly collected fracture image, its size is 1000×4000 pixels. During the training process of the neural network, the crack image size is relatively large for the input of the network. Each image is divided into 100 sub-blocks of 200×200 pixels, a large number of image blocks with cracks ...

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 tunnel lining crack image segmentation method, and belongs to the technical field of image recognition. The method comprises the following steps: firstly, dividing a directlycollected original crack image into a plurality of image blocks, and selecting the image blocks with cracks and corresponding binary images as a training sample set; performing de-mean normalization processing on the image blocks with the cracks in the training sample set; training the training sample set subjected to mean value removal normalization processing by using a convolutional neural network and combining a loss function to obtain a crack segmentation model; and carrying out mean normalization processing on the crack image to be segmented, inputting the crack image to be segmented into the crack segmentation model, and outputting a segmentation result image. According to the method based on the deep convolutional neural network, the image is automatically identified, human intervention is reduced, the influence of subjective factors of a detector is avoided, the identification precision and the identification efficiency are improved, and the false alarm rate is reduced.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method for segmenting tunnel lining crack images. Background technique [0002] The construction process of infrastructure such as roads and railway tunnels is getting faster and faster, so the requirements for safety are also higher, and the demand for perfect protection measures is increasing. Due to weather, construction process and other factors, the inner wall of the tunnel will have different degrees of cracks and other diseases. In the case of severe crack disease, it may cause tunnel collapse accidents. Therefore, regular inspection of diseases in the tunnel environment has a crucial impact on the safe operation of road traffic. [0003] In the past, the repair and protection of tunnels were mainly carried out manually using professional instruments and equipment, which has a high accuracy. However, it is time-consuming and labor-intensive, and the work eff...

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): G06K9/34G06K9/62
CPCG06V10/267G06F18/214
Inventor 李清勇梁凤娇刘滨
Owner 新而锐电子科技(上海)有限公司
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