[chapter 01 image defogging technology based on histogram optimization - full system matlab intelligent driving depth learning]

There is no doubt about it , In automotive intelligent technology 、 Automobile new energy technology 、 Automotive electronics is a race track for the best . The requirements of intelligent driving technology are also constantly improving .
Intelligent vehicle teaching platform 、 Intelligent network teaching platform 、 Automobile electronic teaching equipment 、 In the loop simulation system , At present, there are many mainstream systems , Common basic parts and algorithm development , Intelligent driving module , It is better to get started matlab platform .
MATLAB/Simulink The two key parts of the are ：m The grammar of the language and the various toolkits it carries .

A powerful toolbox covering all areas is MATLAB The core of software , Various function、app、blcoks Are included in these toolkits . therefore MATLAB/Simulink Learning from , The core is the learning of these toolkits . So we are learning MATLAB When , You can learn by using the toolbox as a unit .
and ,matlab It has obvious advantages in resource integration , Great backwardness , Enough to learn the real thing , From entry to mastery

Histogram technology processing

Common means of image enhancement , But the global square equalization often brings distortion to the image , In order to deal with related problems , This paper adopts the global square equilibrium 、 Local square equilibrium and Retinex Algorithm to process the image .
For the specific functions and the application of their syntax , stay matlab Of Help documents can be searched by themselves , Case study and application .
Let's start with an example ：
The effect of processing vehicle tires is shown in the figure .

demo.m

clc; clear all; close all;
I = imread('tire.tif');
J = histeq(I);
figure; 
subplot(2, 2, 1); imshow(I, []); title(' Original picture ');
subplot(2, 2, 2); imshow(J, []); title(' Image after original image equalization ');
subplot(2, 2, 3); imhist(I, 64); title(' Histogram of the original image ');
subplot(2, 2, 4); imhist(J, 64); title(' Histogram after equalization ');

image file , The effect of the original image is shown in the figure

Image defogging technology system based on histogram optimization

GUI The interface is as shown in the figure ：

Image defogging system , First load the image and display , Then select the defogging algorithm , Finally, you can observe the histogram contrast effect

Initfig.m

function InitFig(hObject,handles)
axes(handles.axes1); cla; set(gca, 'Color', [0.8039 0.8784 0.9686]);
axes(handles.axes2); cla; axis on; box on; set(gca, 'Color', [0.8039 0.8784 0.9686]);
set(gca, 'XTickLabel', [], 'YTickLabel', [], 'XTick', [], 'YTick', []);
set(handles.textInfo, 'String', ...
    ' Image defogging system , First load the image and display , Then select the defogging algorithm , Finally, you can observe the histogram contrast effect .');

Screenshot function
SnapImage.m

function SnapImage()
imagesPath = '.\\snap_images';
if ~exist(imagesPath, 'dir')
    mkdir(imagesPath);
end

[FileName,PathName,FilterIndex] = uiputfile({'*.jpg;*.tif;*.png;*.gif','All Image Files';...
          '*.*','All Files' },' Save the screenshot ',...
          '.\\snap_images\\temp.jpg');
if isequal(FileName, 0) || isequal(PathName, 0)
    return;
end
fileStr = fullfile(PathName, FileName);
f = getframe(gcf);
f = frame2im(f);
imwrite(f, fileStr);
msgbox(' The snapshot file was saved successfully ！', ' Prompt information ');

Image saving subfunction ：
SaveImage.m

function SaveImage(Img)
imagesPath = '.\\results';
if ~exist(imagesPath, 'dir')
    mkdir(imagesPath);
end

[FileName,PathName,FilterIndex] = uiputfile({'*.jpg;*.tif;*.png;*.gif','All Image Files';...
          '*.*','All Files' },' Save the screenshot ',...
          '.\\results\\result.jpg');
if isequal(FileName, 0) || isequal(PathName, 0)
    return;
end
fileStr = fullfile(PathName, FileName);
imwrite(Img, fileStr);
msgbox(' Processing results saved successfully ！', ' Prompt information ');

use Retinex Of MSR Realize image defogging operation .
Subfunctions RemoveFogByRetinex.m

function In = RemoveFogByRetinex(f, flag)
%  use Retinex Of MSR Realize image defogging 
%  Input parameters ：
%  f—— Image matrix 
%  flag—— show marks 
%  Output parameters ：
%  In—— The resulting image 

if nargin < 2
    flag = 1;
end
% Extract the R、G、B component 
fr = f(:, :, 1);
fg = f(:, :, 2);
fb = f(:, :, 3);
% Data type normalization 
mr = mat2gray(im2double(fr));
mg = mat2gray(im2double(fg));
mb = mat2gray(im2double(fb));
% Definition alpha Parameters 
alpha = 1200;
% Define template size 
n = 128;
% Computing Center 
n1 = floor((n+1)/2);
for i = 1:n
    for j = 1:n
        % Gauss function 
        b(i,j)  = exp(-((i-n1)^2+(j-n1)^2)/(4*alpha))/(pi*alpha);
    end
end
% Convolution filtering 
nr1 = imfilter(mr,b,'conv', 'replicate');
ng1 = imfilter(mg,b,'conv', 'replicate');
nb1 = imfilter(mb,b,'conv', 'replicate');
ur1 = log(nr1);
ug1 = log(ng1);
ub1 = log(nb1);
tr1 = log(mr);
tg1 = log(mg);
tb1 = log(mb);
yr1 = (tr1-ur1)/3;
yg1 = (tg1-ug1)/3;
yb1 = (tb1-ub1)/3;
% Definition beta Parameters 
beta = 55;
% Define template size 
x = 32;
% Computing Center 
x1 = floor((n+1)/2);
for i = 1:n
    for j = 1:n
        % Gauss function 
        a(i,j)  = exp(-((i-n1)^2+(j-n1)^2)/(4*beta))/(6*pi*beta);
    end
end
% Convolution filtering 
nr2 = imfilter(mr,a,'conv', 'replicate');
ng2 = imfilter(mg,a,'conv', 'replicate');
nb2 = imfilter(mb,a,'conv', 'replicate');
ur2 = log(nr2);
ug2 = log(ng2);
ub2 = log(nb2);
tr2 = log(mr);
tg2 = log(mg);
tb2 = log(mb);
yr2 = (tr2-ur2)/3;
yg2 = (tg2-ug2)/3;
yb2 = (tb2-ub2)/3;
% Definition eta Parameters 
eta = 13944.5;
% Define template size 
l = 500;
% Computing Center 
l1 = floor((n+1)/2);
for i = 1:n
    for j = 1:n
        % Gauss function 
        e(i,j)  = exp(-((i-n1)^2+(j-n1)^2)/(4*eta))/(4*pi*eta);
    end
end
% Convolution filtering 
nr3 = imfilter(mr,e,'conv', 'replicate');
ng3 = imfilter(mg,e,'conv', 'replicate');
nb3 = imfilter(mb,e,'conv', 'replicate');
ur3 = log(nr3);
ug3 = log(ng3);
ub3 = log(nb3);
tr3 = log(mr);
tg3 = log(mg);
tb3 = log(mb);
yr3 = (tr3-ur3)/3;
yg3 = (tg3-ug3)/3;
yb3 = (tb3-ub3)/3;
dr = yr1+yr2+yr3;
dg = yg1+yg2+yg3;
db = yb1+yb2+yb3;
cr = im2uint8(dr);
cg = im2uint8(dg);
cb = im2uint8(db);
%  Integrate the processed components to obtain the result image 
In = cat(3, cr, cg, cb);
% Results show 
if flag
    figure;
    subplot(2, 2, 1); imshow(f); title(' Original image ', 'FontWeight', 'Bold');
    subplot(2, 2, 2); imshow(In); title(' Processed image ', 'FontWeight', 'Bold');
    %  Graying , Used to calculate histogram 
    Q = rgb2gray(f);
    M = rgb2gray(In);
    subplot(2, 2, 3); imhist(Q, 64); title(' Original gray histogram ', 'FontWeight', 'Bold');
    subplot(2, 2, 4); imhist(M, 64); title(' Processed gray histogram ', 'FontWeight', 'Bold');
end

Gray histogram and image normalization function ：
RemoveFogByGlobalHisteq.m

function In = RemoveFogByGlobalHisteq(I, flag)

if nargin < 2
    flag = 1;
end
R = I(:,:,1);
G = I(:,:,2);
B = I(:,:,3);
M = histeq(R);
N = histeq(G);
L = histeq(B);
In = cat(3, M, N, L);
if flag
    figure;
    subplot(2, 2, 1); imshow(I); title(' Original image ', 'FontWeight', 'Bold');
    subplot(2, 2, 2); imshow(In); title(' Processed image ', 'FontWeight', 'Bold');
    Q = rgb2gray(I);
    W = rgb2gray(In);
    subplot(2, 2, 3); imhist(Q, 64); title(' Original gray histogram ', 'FontWeight', 'Bold');
    subplot(2, 2, 4); imhist(W, 64); title(' Processed gray histogram ', 'FontWeight', 'Bold');
end

All the source code of this article is packaged and sorted out –> Portal

Image defogging technology based on histogram optimization - The whole system –MATLAB Intelligent driving deep learning
The global processing effect is shown in the figure .

The effect of local histogram equalization

adopt Retinex Comparison of the fog removal effect of the algorithm ：

Compare with the corresponding histogram

Run the main process file
MainForm.fig
Download a full set of source code