Unlock instant, AI-driven research and patent intelligence for your innovation.

Structural surface crack detection method under small sample based on generative adversarial network

A technology of structural surfaces and detection methods, applied in biological neural network models, neural learning methods, neural architectures, etc., can solve problems such as accuracy decline, and achieve the effect of improving performance

Pending Publication Date: 2022-03-01
TONGJI UNIV
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] However, the above existing related studies did not consider the case where the data set is too small to train the model. In fact, if the data set used for training is small, the trained model is used to detect pavement cracks. will drop significantly

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
  • Structural surface crack detection method under small sample based on generative adversarial network
  • Structural surface crack detection method under small sample based on generative adversarial network
  • Structural surface crack detection method under small sample based on generative adversarial network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0096] The structural surface crack detection method in this example includes the following steps:

[0097] Step 1, the collection of data sets in one road of two road intersections, using drones, drone data is shown in Table 4:

[0098] Table 4 drone parameters

[0099] parameter value UAV mass / g 1250 Maximum flight time / min About 30 Effective Pixels About 20,000,000 Photo size 5472×3648 Video resolution 1920×1080 Controlled rotation range / (°) -90°-30°

[0100] The drone flying height is 1.5m, the camera is perpendicular to the ground, the drone is moving in the road direction, capturing an image of 80 ways, first pretreatment. The picture pixel size taken by the drone is 5472 × 3648, since the convolutional neural network treatment is generally a square image, the image is divided into images of 24 dimensions 912 × 912.

[0101] The image compressed by the divided resolution is 912 × 912 is an image of 224 × 224, and finally the ...

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 small sample structure surface crack detection method based on a generative adversarial network. The method comprises the following steps: 1) acquiring an image of a structure surface and preprocessing the image to construct an image data set; 2) constructing an image generation model to train the collected image data set to realize expansion of the image data set; and 3) constructing a detection model for judging whether a crack exists on the surface of the structure based on the VggNet, and training to complete crack detection on the surface of the structure. Compared with the prior art, the method has the advantages of being suitable for small samples, accurate and rapid in detection and the like.

Description

Technical field [0001] The present invention relates to the field of structural surface crack detection, and more particularly to a method of structural surface crack detection method based on a generating type counterfeit network. Background technique [0002] In recent years, with the large number of construction and use of the road, the maintenance and management of roads will gradually be concerned. In the early days of the road, the road detection and maintenance can be greatly reduced. With the rapid development of deep learning technology, the collected structural surface images can be quickly and accurately classified and detected using a convolved neural network. For tasks based on structural surface detection of convolutional neural networks, how to obtain a large number of high quality training set data is a key issue that needs to be solved when the number of structural surfaces of the convolutional neural network is missing. [0003] The research on the prior art is ...

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): G06N3/04G06N3/08G06K9/62G06V20/60G06V10/774G06V10/82
CPCG06N3/084G06N3/045G06F18/2414G06F18/214
Inventor 刘超许博强
Owner TONGJI UNIV