Matlab实现遗传算法在二值图像重建问题中的应用

小咪咪 2022-08-30 00:48 202阅读 0赞

# 遗传算法 #

GA 算法的灵感来自查尔斯·达尔维 (Charles Darwi) 提出的生物进化理论。实际上，GA中模拟的主要机制是适者生存。在自然界中，更健康的生物体生存的可能性更高。这有助于他们将基因转移到下一代。随着时间的推移，让物种更好地适应环境（避开敌人并寻找食物）的良好基因在后代中占据主导地位。受自然界中染色体和基因的启发，GA 算法将优化问题表示为一组变量。优化问题的每个解对应一个染色体，每个基因对应问题的一个变量。

进化计算

进化计算是一系列受生物进化启发的全局优化算法，是研究这些算法的人工智能和软计算的子领域。从技术上讲，它们是一系列具有元启发式或随机优化特征的基于群体的试错问题求解器。

本文是将遗传算法用于图像重建

代码：

function [BestChrom]  = GeneticAlgorithm (M , N, MaxGen , Pc, Pm , Er , obj, visuailzation)
    
    load IMG_REF_BINARY;
    
    cgcurve = zeros(1 , MaxGen);
    
    %%  Initialization
    [ population ] = initialization(M, N);
    for i = 1 : M
        population.Chromosomes(i).fitness = obj( population.Chromosomes(i).Gene(:), IMG_REF_BINARY);
    end
    
    g = 1;
    disp(['Generation #' , num2str(g)]);
    [max_val , indx] = sort([ population.Chromosomes(:).fitness ] , 'descend');
    cgcurve(g) = population.Chromosomes(indx(1)).fitness;
    
    subplot(1,2,2)
    
    %% Main loop
    for g = 2 : MaxGen
        disp(['Generation #' , num2str(g)]);
    
        for k = 1: 2: M
            % Selection
            [ parent1, parent2] = selection(population);
            
            % Crossover
            [child1 , child2] = crossover(parent1 , parent2, Pc, 'double');
            
            % Mutation
            [child1] = mutation(child1, Pm);
            [child2] = mutation(child2, Pm);
            
            newPopulation.Chromosomes(k).Gene = child1.Gene;
            newPopulation.Chromosomes(k+1).Gene = child2.Gene;
        end
        
        % Calcualte the fitness values
        for i = 1 : M
            newPopulation.Chromosomes(i).fitness = obj( newPopulation.Chromosomes(i).Gene(:), IMG_REF_BINARY);
        end
        
        % Elitism
        [ newPopulation ] = elitism(population, newPopulation, Er);
        
        cgcurve(g) = newPopulation.Chromosomes(1).fitness;
        
        population = newPopulation; % Replace the previous population with the newly made
        
        BestChrom.Gene    = population.Chromosomes(1).Gene;
        BestChrom.Fitness = population.Chromosomes(1).fitness;
        
        
        col_no = size(IMG_REF_BINARY,2);
        Recons_IMG = vec2mat(BestChrom.Gene , col_no);
        Recons_IMG = Recons_IMG .* 255;
    
        
     %   if rem(g , 100) == 0
            subplot(1,2,2)
            imshow(Recons_IMG);
            title(['Generation #' , num2str(g), ' Error = ' , num2str(-BestChrom.Fitness)])
     
            drawnow
     %   end
        
    end
    
       
    
    
    
    if visuailzation == 1
        figure
        plot( 1 : MaxGen , cgcurve);
        xlabel('Generation');
        ylabel('Fitness of the best elite')
    end
    
    
    end

运行结果：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3VwZGF0ZTc_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3VwZGF0ZTc_size_16_color_FFFFFF_t_70 1][]

获取完整代码：[https://ai.52learn.online/code/40][https_ai.52learn.online_code_40]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3VwZGF0ZTc_size_16_color_FFFFFF_t_70]: /images/20220829/596edf21a05748528107d676939c8979.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3VwZGF0ZTc_size_16_color_FFFFFF_t_70 1]: /images/20220829/337ff260fe8c4710a9286f1ea0960e3f.png
[https_ai.52learn.online_code_40]: https://ai.52learn.online/code/40