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

Pavement defect detection method and system based on deep learning

A defect detection and deep learning technology, which is applied in neural learning methods, image data processing, image enhancement, etc., can solve the problems of single feature and reduce the amount of calculation, so as to reduce the amount of calculation, remove redundant information, and improve detection efficiency Effect

Pending Publication Date: 2022-02-18
INNER MONGOLIA UNIV OF SCI & TECH
View PDF0 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method and system for pavement defect detection based on deep learning. By setting U-MDM, the module adopts a four-scale up-down sampling structure, and combines the advantages of dilated convolution to overcome the commonly used standard convolution. The shortcoming of a single feature extracted by the product, and redundant information can be removed to reduce the amount of calculation

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
  • Pavement defect detection method and system based on deep learning
  • Pavement defect detection method and system based on deep learning
  • Pavement defect detection method and system based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0041] The U-MDN designed for the present invention is composed of two parts of U-NET and U-MDM in the overall network structure. First use U-NET to extract multi-scale information from different convolutional blocks, and then use U-MDM to further extract deep features from four different sizes, and finally simultaneously fuse the shallow feature with deep features using U-NET. Finally, the prediction result is obtained through the Sigmoid function.

[0042] Division network design, U-NET depth neural networks are symmetrical U-shaped structures containing the lower sampling image compression path and the upper sampling image extension path. This structure is a basic idea of ​​the FCN codec structure. And both networks use convolution operation instead of full connection operation, which greatly reduces the amount of network model parameters and improves the model training speed. Different, U-NET will focus on the upper sampling stage, increasing the number of Feature Map in the a...

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 pavement defect detection method and system based on deep learning, wherein the method comprises the following steps: building an image segmentation model, taking a U-Net feature extraction network as a trunk by U-MDN, extracting multi-scale features obtained from different convolution stages, further extracting defect deep layer features by using U-MDM, efficiently fusing multi-scale features obtained in shallow layer features and the deep layer features to obtain a prediction result; and carrying out pavement defect quantitative evaluation: further calculating a pavement damage index based on the quantitative parameters to obtain a corresponding pavement damage grade. According to the method, the defect of single feature of common standard convolution extraction is overcome, redundant information can be removed, and the calculation amount can be reduced; and the pavement defects are detected in real time by using the improved deep neural network, the defects of a traditional manual pavement detection method are overcome, and the detection efficiency is improved.

Description

Technical field [0001] The present invention relates to the field of pavement defect detection, and more particularly to a depth study-based pavement defect detection method and system. Background technique [0002] With the rapid development of the transportation industry, my country has formed a road traffic network with a certain scale. A large number of surface defects generated during use marks the existing roads have entered a large area detection and maintenance phase. The more and more diversification and complexity of highway traffic networks have caused the performance requirements of the defect detection method to be higher and higher. Therefore, it is of great significance to find and more accurately and quickly achieve crack detection on road maintenance and driving safety. [0003] Pavement cracks are one of the most common diseases, which has a huge threat to road safety. It is possible to find that repair and prevent further deterioration has become an important t...

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): G06T7/00G06T7/11G06T7/62G06T7/66G06N3/08G06N3/04G06K9/62G06V10/82G06V10/774
CPCG06T7/0004G06N3/08G06T7/11G06T7/62G06T7/66G06T2207/20084G06T2207/20081G06N3/045G06F18/253
Inventor 陈波李琦王鑫白卓玉于令君赵建敏王月明
Owner INNER MONGOLIA UNIV OF SCI & TECH
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