Method for segmenting color images on basis of NSGA-II (non-dominated sorting genetic algorithm-II) evolution algorithms

Method for segmenting color images on basis of NSGA-II (non-dominated sorting genetic algorithm-II) evolution algorithms

  • CN 104,952,067 A
  • Filed: 05/13/2015
  • Published: 09/30/2015
  • Est. Priority Date: 05/13/2015
  • Status: Active Application
First Claim
Patent Images

1. , based on a color image segmentation method for NSGA-II evolution algorithm, comprise the steps:

  • (1) image to be split is inputted;

    (2) mark cluster are carried out to the pixel in image to be split, obtain marking and image after cluster;

    (3) use dividing ridge method to carry out coarse segmentation to the image after mark also cluster, obtain the image after coarse segmentation;

    (4) initialization of population is carried out to the image after coarse segmentation, obtain initialization population;

    (5) fitness of the individuality in initialization population is calculated;

    (6) initialization population non-dominant quicksort, obtains sequence number and the crowding distance of sequence;

    6a) utilize the non-dominant quicksort in NSGA-II algorithm to sort to the individuality in initialization population, generate the sequence number of sequence, the individuality in initialization population according to sequence number divided rank from small to large;

    Crowding distance between individuality 6b) utilizing the crowding distance calculating method in NSGA-II algorithm to calculate in same grade;

    (7) parent population is selected according to the sequence number sorted and crowding distance by binary tournament algorithm;

    (8) parent population crossover and mutation generates filial generation;

    8a) interlace operation is done to parent population, obtain the progeny population after intersecting;

    8b) mutation operation is done to parent population, obtain the progeny population after making a variation;

    (9) non-dominant quicksort is carried out in parent and progeny population merging, obtains sequence number and the crowding distance of sequence;

    Individuality 9a) utilizing the non-dominant quicksort in NSGA-II algorithm to be combined in rear population sorts, and generates the sequence number of sequence, according to sequence number from small to large the individual divided rank merged in rear population;

    Crowding distance between individuality 9b) utilizing the crowding distance calculating method in NSGA-II algorithm to calculate in same grade;

    (10) elite population is selected according to the sequence number sorted and crowding distance by binary tournament algorithm;

    (11) judge whether to reach cyclic algebra setting model, if do not reach, jump to step (7), and circulation is until reach cyclic algebra setting model, export pareto front end;

    (12) from pareto front end, choose a solution, to the pixel assignment of image to be split, obtain the final segmentation result of all pixels in image to be split.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×