Underwater sonar image unsupervised classification method based on class consciousness field adaptation

An underwater sonar and classification method technology, applied in image analysis, image enhancement, image data processing and other directions, can solve problems affecting underwater sonar image classification research, etc., to enhance the feature extraction ability, increase the distance between classes, The effect of improving accuracy

Active Publication Date: 2020-07-24
HARBIN ENG UNIV
View PDF5 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This also greatly affects the classifica

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 sonar image unsupervised classification method based on class consciousness field adaptation
  • Underwater sonar image unsupervised classification method based on class consciousness field adaptation
  • Underwater sonar image unsupervised classification method based on class consciousness field adaptation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0033] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0034] The present invention comprises the following steps in the realization process:

[0035] (1) Use CGAN and DCGAN to generate underwater sonar images to construct generated datasets, and construct two original datasets of underwater sonar images, balanced and unbalanced;

[0036] (2) An improved method based on AAE is proposed to build a domain-adaptive source domain: ① extract features and perform semantic segmentation; ② use pseudo-label technology to train the source domain model on the underwater sonar image generation dataset; ③ transfer the pseudo-label to Input the improved AAE model after hot encoding format to extract source domain features;

[0037] (3) A method based on adversarial learning is proposed to construct a domain-adaptive target domain: use the source domain model parameters to initialize the target domain ...

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 sonar image unsupervised classification method based on class consciousness domain self-adaptation. The method comprises the steps of (1) using a generative adversarial network for constructing and generating a data set; (2) proposing an improved method based on an adversarial auto-encoder to construct a domain-adaptive source domain; (3) proposing to constructa domain-adaptive target domain based on an adversarial learning method; and (4) training a target domain, and completing unsupervised classification of the underwater sonar images on the balanced native data set and the non-balanced native data set. According to the invention, two GANs of CGAN and DCGAN are used to generate images so as to construct an underwater sonar image generation data set,and an unsupervised domain adaptive method is introduced into unsupervised classification of underwater sonar images according to the situation of label missing. And a balanced underwater sonar imagenative data set and an unbalanced underwater sonar image native data set are simultaneously constructed to verify the adaptability of the method provided by the invention.

Description

technical field [0001] The invention relates to an unsupervised classification method for underwater sonar images, in particular to an adaptive unsupervised classification method for underwater sonar images based on class awareness domain, which belongs to the field of underwater sonar image classification. Background technique [0002] Since the country put forward the strategy of developing a marine power, new requirements have been put forward for ocean exploration, resource utilization, and marine technology and equipment. How to more accurately discover marine resources is the premise of marine exploration and utilization, and since marine resources are mainly natural gas, marine organisms, and ship remains, relevant underwater detection and identification technologies have become the key to marine exploration and utilization technology. Mastering and innovating underwater resource classification technology is the premise of detection and identification, therefore, und...

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): G06K9/62G06K9/34G06K9/46G06T7/11
CPCG06T7/11G06T2207/10004G06V10/267G06V10/40G06F18/214G06F18/241
Inventor 王兴梅孙博轩王坤华徐义超孟稼祥
Owner HARBIN ENG 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