三维重建之多频外差解包裹学习笔记
Posted u012507022
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了三维重建之多频外差解包裹学习笔记相关的知识,希望对你有一定的参考价值。
附matlab多频外差解相位程序
%参考博客:https://blog.csdn.net/qq_15295565/article/details/99704919,并进行修改
clc;
close all;
clear;
%% 初始化
width = 1024;
heigth = 768;
%三频四步
FREQ = 3;
STEP = 4;
%freq为三个正弦条纹对应的周期数,λ=width/freq:代表每个条纹的周期(波长)
% 也可理解为频率,三个频率参考李中伟的博士论文
freq = [70 64 59];
freq12 = 6 ;
freq23 = 5 ;
freq123 = 1 ;
%% Step0 生成条纹图
% 利用分块矩阵C存储3组共计12张图,三种频率,四步相位
C = cell(FREQ,STEP);
for i=1:FREQ
for j=1:STEP
Ci,j = zeros(heigth,width);
end
end
% 利用余弦函数计算12张图的灰度值,图像的生成
% 三种频率,四组相位
for i = 1:FREQ % 对应三种频率
for j = 0:(STEP) % 对应四步相位
for k = 1:width
Ci,j+1(:,k) = 128+127*sin(2*pi*k*freq(i)/width+j*pi/2);
end
end
end
% 归一化处理
for i = 1:3
for j = 1:4
Ci,j = mat2gray(Ci,j);
end
end
% 显示12张图
% for i = 1:FREQ
% for j = 1:STEP
% n = STEP*(i-1)+j;
% h = figure(n);
% imshow(Ci,j);
% end
% end
%% Step1 求解相位主值
% phi存储相位主值图像
phi = cell(FREQ,1);
for i = 1:FREQ
phii,1 = zeros(heigth,width);
end
% 计算每种频率对应的相位主值
% 输出三种频率的相位主值,用于相差计算
for i = 1:FREQ % 对于3组中的每一组图片,每一组相同频率的有四张图片
I1 = Ci,1;
I2 = Ci,2;
I3 = Ci,3;
I4 = Ci,4;
for g = 1:heigth
for k = 1:width
if I4(g,k)==I2(g,k)&&I1(g,k)>I3(g,k) %四个特殊位置
phii,1(g,k)=0;
elseif I4(g,k)==I2(g,k)&&I1(g,k)<I3(g,k) %四个特殊位置
phii,1(g,k)=pi;
elseif I1(g,k)==I3(g,k)&&I4(g,k)>I2(g,k) %四个特殊位置
phii,1(g,k)=pi/2;
elseif I1(g,k)==I3(g,k)&&I4(g,k)<I2(g,k) %四个特殊位置
phii,1(g,k)=3*pi/2;
elseif I1(g,k)<I3(g,k) %二三象限
phii,1(g,k)=atan((I4(g,k)-I2(g,k))./(I1(g,k)-I3(g,k)))+pi;
elseif I1(g,k)>I3(g,k)&&I4(g,k)>I2(g,k) %第一象限
phii,1(g,k)=atan((I4(g,k)-I2(g,k))./(I1(g,k)-I3(g,k)));
elseif I1(g,k)>I3(g,k)&&I4(g,k)<I2(g,k) %第四象限
phii,1(g,k)=atan((I4(g,k)-I2(g,k))./(I1(g,k)-I3(g,k)))+2*pi;
end %endif
end
end
end
% 显示
% figure,imshow(mat2gray(PH1));title('1相位主值');
% figure,imshow(mat2gray(PH2));title('2相位主值');
% figure,imshow(mat2gray(PH3));title('3相位主值');
%% Step2 利用外差原理叠加不同频率的相位主值,得到整幅图像只有一个周期的相位
PH1 = phi1,1; %频率1
PH2 = phi2,1; %频率2
PH3 = phi3,1; %频率3
PH12 = zeros(heigth,width);
PH23 = zeros(heigth,width);
PH123 = zeros(heigth,width);
for g = 1:heigth
for k = 1:width
% 计算第1组和第2组的相差
if PH1(g,k)>PH2(g,k)
PH12(g,k) = PH1(g,k)-PH2(g,k);
else
PH12(g,k) = PH1(g,k)+2*pi-PH2(g,k);
end
% 计算第2组和第3组的相差
if PH2(g,k)>PH3(g,k)
PH23(g,k) = PH2(g,k)-PH3(g,k);
else
PH23(g,k) = PH2(g,k)+2*pi-PH3(g,k);
end
% 计算第12组和第23组的相差
if PH12(g,k)>PH23(g,k)
PH123(g,k) = PH12(g,k)-PH23(g,k);
else
PH123(g,k) = PH12(g,k)+2*pi-PH23(g,k);
end
end
end
%显示
figure,imshow(mat2gray(PH12));title('1,2外差');
figure,imshow(mat2gray(PH23));title('2,3外差');
figure,imshow(mat2gray(PH123));title('1,2,3外差');
%% Step3 相位解包裹;
%根据公式:
%η = floor(((delta_φ*delta_λ/λ)-φ)/2π),
%Φ = 2πη+φ ,
%由PH123、PH23、PH12反计算出PH1、PH2、PH3的连续相位
%初始化解包裹相位
phUnWrap = cell(3,1);
for i = 1:3
phUnWrapi,1 = zeros(heigth,width);
end
% 解包裹23:求PH23的绝对相位,deltaPh = PH123,phWrap = PH23
deltaPh = PH123;
phWrap = PH23;
R = freq23/freq123;
Nwrap = floor((deltaPh*R-phWrap)/(2*pi)+0.5);
ph23UnWrap= (2*pi)*Nwrap +phWrap;
% 解包裹12:求PH12的绝对相位,deltaPh = PH123,phWrap = PH12
deltaPh = PH123;
phWrap = PH12;
R = freq12/freq123;
Nwrap = floor((deltaPh*R-phWrap)/(2*pi)+0.5);
ph12UnWrap= (2*pi)*Nwrap +phWrap;
% 解包裹1:求PH1的绝对相位,deltaPh = PH12,phWrap = PH1
deltaPh = ph12UnWrap;
phWrap = PH1;
R = freq(1)/freq12;
Nwrap = floor((deltaPh*R-phWrap)/(2*pi)+0.5);
phUnWrap1,1= (2*pi)*Nwrap +phWrap;
% 解包裹2:求PH2的绝对相位,deltaPh = PH12,phWrap = PH2
deltaPh = ph12UnWrap;
phWrap = PH2;
R = freq(2)/freq12;
Nwrap = floor((deltaPh*R-phWrap)/(2*pi)+0.5);
phUnWrap2,1= (2*pi)*Nwrap +phWrap;
% 解包裹3:求PH2的绝对相位,deltaPh = PH23,phWrap = PH3
deltaPh = ph23UnWrap;
phWrap = PH3;
R = freq(3)/freq23;
Nwrap = floor((deltaPh*R-phWrap)/(2*pi)+0.5);
phUnWrap3,1= (2*pi)*Nwrap +phWrap;
%% show phUnWrap
figure
plot(PH123(100,:),'r');hold on
plot(PH23(100,:),'g');hold on
plot(ph23UnWrap(100,:));
grid on
legend('deltaPh','wrapPH','phUnWrap');
title('解包裹23');
figure
plot(PH123(100,:),'r');hold on
plot(PH12(100,:),'g');hold on
plot(ph12UnWrap(100,:));
grid on
legend('deltaPh','wrapPH','phUnWrap');
title('解包裹12');
figure
plot(ph12UnWrap(100,:),'r');hold on
plot(PH1(100,:),'g');hold on
plot(phUnWrap1,1(100,:));
grid on
legend('deltaPh','wrapPH','phUnWrap');
title('解包裹1');
figure
plot(ph12UnWrap(100,:),'r');hold on
plot(PH2(100,:),'g');hold on
plot(phUnWrap2,1(100,:));
grid on
legend('deltaPh','wrapPH','phUnWrap');
title('解包裹2');
figure
plot(ph23UnWrap(100,:),'r');hold on
plot(PH3(100,:),'g');hold on
plot(phUnWrap3,1(100,:));
grid on
legend('deltaPh','wrapPH','phUnWrap');
title('解包裹3');
参考:
https://blog.csdn.net/qq_15295565/article/details/99704919
https://blog.csdn.net/gao_summer_cola/article/details/75308618
https://mp.weixin.qq.com/s/eTlkyNrqDk_qrb7eS91ktQ
https://zhuanlan.zhihu.com/p/82521153
以上是关于三维重建之多频外差解包裹学习笔记的主要内容,如果未能解决你的问题,请参考以下文章