Underwater pier component segmentation method based on deep learning and sonar imaging

A deep learning and sonar technology, applied in image analysis, neural architecture, image enhancement, etc., can solve the problems of low automation, easy misjudgment, high cost, etc., and achieve high automation, low cost, and high efficiency Effect

Inactive Publication Date: 2020-02-28
SOUTHEAST UNIV
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The underwater part of the bridge pier has been subjected to harsh environments such as erosion and corrosion for a long time, which may cause defects such as defects, cracks, and exposed ribs or even damage underwater, which will seriously affect the service life and even the bearing capacity of the bridge.
At present, the detection methods of underwater bridge piers are still based on artificial div...

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
  • Underwater pier component segmentation method based on deep learning and sonar imaging
  • Underwater pier component segmentation method based on deep learning and sonar imaging
  • Underwater pier component segmentation method based on deep learning and sonar imaging

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] The present invention will be described in further detail below in conjunction with the accompanying drawings. Such as figure 1 As shown, a method for segmenting underwater pier components based on deep learning and sonar imaging includes the following steps:

[0023] 1. Use side-sound sonar equipment to obtain pictures of underwater piers. The original pictures of sonar are as follows: image 3 , Figure 4 As shown, form a data set and set each picture to a size of 1200x1200 pixel.

[0024] 2. Use the method of data enhancement to expand the data set, and label each picture. The data enhancement method here uses random rotation and random cropping, and the rotation angle is 90 degrees and 180 degrees. The label is marked with a polygonal frame, using LabelImg Software annotations.

[0025] 3. Divide the data set into training test set, verification set and test set according to the ratio of 8:1:1.

[0026] 4. Establish a deep learning MaskRCNN model, such as fig...

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 an underwater pier component segmentation method based on deep learning and sonar imaging. The underwater pier component segmentation method comprises: using underwater side sonar equipment for obtaining a prepared underwater pier scanning picture; increasing the number of data sets by using an image enhancement method; marking the data set, carrying out polygon marking on the pier, the pile foundation and the riverbed by using different colors, and recording polygon vertex coordinates; dividing the data set into a training and verification set: establishing a Mask RCNNmodel in a deep learning semantic segmentation network, and performing training to obtain a training model; and controlling side sonar equipment to scan along the underwater pier part on the water surface to obtain a scanning picture, and performing automatic component segmentation on the underwater pier by utilizing the trained Mask RCNN model. The method is high in efficiency and low in cost, and compared with a traditional manual diving method and a sonar manual screening method, the method has the obvious advantages of automation, high efficiency and accuracy.

Description

technical field [0001] The invention belongs to the technical field of interaction between civil engineering and artificial intelligence, and in particular relates to a method for segmenting underwater pier components based on deep learning and sonar imaging. Background technique [0002] The bridge pier is the main load-bearing component of the bridge pier, and most of the load of the bridge structure is transmitted to the foundation through the pier. Any loss of bearing capacity of a bridge pier will lead to the overall instability and destruction of the bridge pier. Therefore, the safety of the bridge pier must be highly valued. The underwater part of the bridge pier has been subjected to harsh environments such as erosion and corrosion for a long time, which may cause defects such as defects, cracks, and exposed ribs or even damage underwater, which will seriously affect the service life and even the bearing capacity of the bridge. At present, the detection methods of u...

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): G06T7/10G06T5/00G06N3/04
CPCG06T7/10G06T5/002G06N3/045
Inventor 侯士通吴刚董斌
Owner SOUTHEAST 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