Tunnel crack identification method based on deep learning and OpenCV

A deep learning and crack recognition technology, applied in the field of deep learning and graphics processing, can solve the problems of high accuracy of labeling data, difficulty in labeling, and few samples.

Active Publication Date: 2021-03-05
SHANDONG UNIV +1
View PDF8 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the network requires a large amount of labeled data, and the accuracy of the labeled data is high, that is, the requirements for the data set are relatively strict, and a large amount of accurately labeled training data is required for target objects with complex features, while cracks, especially tunnels Cracks, due to

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 crack identification method based on deep learning and OpenCV
  • Tunnel crack identification method based on deep learning and OpenCV
  • Tunnel crack identification method based on deep learning and OpenCV

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] A tunnel crack recognition method based on deep learning and OpenCV, including:

[0074] Step 1: Use deep learning technology to find out the general location and shape of the crack: based on Mask-RCNN, figure 1 It is a schematic diagram of the Mask-RCNN network structure. Using the labeled tunnel crack picture data set to train a model that can identify the general position and shape of the crack, a mask of a crack picture is given, such as figure 2 As shown, the next step is to extract cracks within the mask range;

[0075] Step 2: Use image processing technology to refine the identified cracks: first, perform corrosion operations on the mask to make it contain more comprehensive crack information, such as image 3 As shown, the skeleton is extracted in the corroded mask area; then the extracted crack skeleton is combined with the original image, and the crack is filled by the neighborhood region growing algorithm;

[0076] Step 3: Statistics of crack length and w...

Embodiment 2

[0078] A method for identifying tunnel cracks based on deep learning and OpenCV, as shown in Embodiment 1, the difference is that step 1 specifically includes the following steps:

[0079] 1.1. Input the entire image into the network;

[0080] 1.2. Input the image into CNN for feature extraction. In this layer, the masks obtained by different scales are finally superimposed as the mask of the original image;

[0081] 1.3. Use FPN to generate proposal windows (proposals), and generate N proposal windows for each picture;

[0082] 1.4. Map multiple suggestion windows generated by FPN to the last layer of convolution feature map of CNN;

[0083] 1.5. Through the RoI Align layer, each RoI generates a fixed-size feature map;

[0084] 1.6. Finally, use full connection classification, border, and mask for regression.

[0085] exist figure 1 Among them, conv is a convolutional layer for feature extraction; RPN is a region generation network for extracting candidate boxes; L is a ...

Embodiment 3

[0088] A tunnel crack recognition method based on deep learning and OpenCV, as shown in Embodiment 1, the difference is that the skeleton extraction process in step 2 is:

[0089] 2.1. Image grayscale: use the weighted average method to weight the values ​​of R, G, and B according to the weights of R: 0.299, G: 0.587, and B: 0.114;

[0090] 2.2. Contrast enhancement:

[0091] The characteristic parameters of the picture include the overall gray value and the overall gray value variance. The histogram is divided into two parts by using the OTSU algorithm. The gray value of the former part is small, and the gray value of the latter part is large. , which may be the gray value of the crack, with few features and small gray value, and the latter part is defined as the background gray value, which may be the gray value of the background, with many features and large gray value, calculate the gray value of these two parts respectively The mean value of the gray level, the differenc...

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 tunnel crack recognition method based on deep learning and OpenCV, and belongs to the technical field of deep learning and graphic processing, and the method comprises the steps: the general position and shape of a crack are found through the deep learning technology; an image processing technology is used to refine an identified crack. Firstly, corrosion operation is performed on mask, and skeleton extraction is performed in a corroded mask area; then, the extracted crack skeleton is combined with an original image, and cracks are filled with an intra-neighborhood region growth algorithm; and crack length and width information are counted. According to the invention, a deep learning technology and a traditional image processing technology are combined, and cracks are accurately extracted in a mask region obtained by an instance segmentation network Mask-RCNN, so that the defect that a result obtained by deep learning is inaccurate and the defect that a result obtained by a classic image processing algorithm is incomplete are overcome; and after the deep learning technology and the traditional image processing technology are combined, accurate and complete cracks can be extracted on one graph.

Description

technical field [0001] The invention relates to a tunnel crack recognition method based on deep learning and OpenCV, and belongs to the technical field of deep learning and graphic processing. Background technique [0002] The current existing instance segmentation network Mask-RCNN can find and accurately segment objects in the image to a certain extent. However, the network requires a large amount of labeled data, and the accuracy of the labeled data is high, that is, the requirements for the data set are relatively strict, and a large amount of accurately labeled training data is required for target objects with complex features, while cracks, especially tunnels Cracks, due to the difficulty of labeling themselves, it is impossible to accurately label every picture, and the features are not obvious, and there are few existing samples, so the segmentation results obtained after training are not ideal. [0003] Traditional image processing algorithms are difficult to accur...

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
IPC IPC(8): G06T7/00G06T7/13G06T7/62G06T7/187
CPCG06T7/0002G06T7/62G06T7/13G06T7/187G06T2207/20081G06T2207/20084
Inventor 刘健韩勃吕高航左志武王凯王剑宏解全一金岩常洪雷
Owner SHANDONG UNIV
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