• DocumentCode
    2766473
  • Title

    Froth image segmentation algorithm

  • Author

    Wang, Luya ; Wang, Weixing

  • Author_Institution
    Inst. of Image Recognition & Comput. Vision, Hunan Normal Univ., Changsha, China
  • Volume
    1
  • fYear
    2000
  • fDate
    2000
  • Firstpage
    480
  • Abstract
    This paper describes a froth image segmentation algorithm. In order to determine the bubble size, shape and texture information from the froth images, the images have to be segmented. Because of the large variation of the froth image patterns and quality, it is difficult to use the existing segmentation algorithm. Therefore, we modified and combined different existing image segmentation algorithms to form a new segmentation algorithm for froth images. In the algorithm, we used and modified a thresholding algorithm to automatically detect the maximum bubbles, then perform a morphological technique to growing bubble regions. The over-segmentation problem was resolved by an object merging algorithm based on bubble shape analysis. In the paper, three froth image segmentation results are demonstrated, which show that the developed segmentation algorithm is reasonable for froth images
  • Keywords
    bubbles; feature extraction; image segmentation; image texture; mathematical morphology; bubble regions; bubble shape analysis; bubble size; froth image patterns; froth image quality; froth image segmentation algorithm; machine vision system; morphological technique; object merging algorithm; over-segmentation problem; shape information; texture information; thresholding algorithm; Algorithm design and analysis; Computer aided manufacturing; Computer vision; Image segmentation; Inspection; Machine vision; Merging; Optical computing; Production; Shape;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing Proceedings, 2000. WCCC-ICSP 2000. 5th International Conference on
  • Conference_Location
    Beijing
  • Print_ISBN
    0-7803-5747-7
  • Type

    conf

  • DOI
    10.1109/ICOSP.2000.894536
  • Filename
    894536