+Advanced Search

Underwater multi-object segmentation technology based on spectral clustering with multi-feature weighting

1.School of Engineering,Dali University;2.College of Physics and Optoelectronic Engineering,Harbin Engineering University;3.School of Electrical and Information Technology,Yunnan Minzu University

Fund Project:

Natural Science Foundation of China (61761048);Special Basic Cooperative Research Programs of Yunnan Provincial Undergraduate Universities’Association [2019FH001(-066)]; Natural Science Foundation of Heilongjiang Province ( LC2018026)

  • Article
  • |
  • Figures
  • |
  • Metrics
  • |
  • Reference
  • |
  • Related
  • |
  • Cited by
  • |
  • Materials

    Sonar image is seriously polluted by noise, which leads to the problem of low precision in underwater multi-target segmentation.Therefore, an underwater multi-object segmentation technique based on self-adjusting spectrum clustering combined with entropy weight method is proposed.The technology first by self-tuning spectral clustering of sonar image pixel clustering processing, make the image is divided into multiple independent area, and then according to the characteristics of complementarity and more sections of the redundancy of the statistical information entropy characteristics, brightness, contrast, long and narrow degree, entropy weight method is used to analyse the characteristics more empowerment and the optimal selection of a target area,Then the optimal target region is matched with all regions by multi-feature similarity. Finally, all target regions are segmented automatically by adaptive threshold iterative method according to the matching results of similarity. Experimental results show that there is not over-segmented of noise interference regions, and target regions segmented have higher accuracy, which verifies the effectiveness of the proposed method.

    Cited by
Article Metrics
  • PDF:
  • HTML:
  • Abstract:
  • Cited by:
Get Citation
  • Received:November 24,2021
  • Revised:March 18,2022
  • Adopted:March 21,2022
  • Online: May 17,2022
  • Published: