利用模板法的光条中心提取

算法描述

线结构光的条纹形状随着物体表面的变化而不同，在比较小的范围内，可 以认为线结构光条纹只有四种模式，水平，垂直，左倾45°，右倾45°。为了能够准确地描述线结构光条纹的形状，在模块选取时一般根据光带的粗细程度进行选取，因为太大不能准确的反映出条纹的细节，太小又不能反映出光带的走向，因此选取了一个9x9的模块来与之相匹配对应这四种模式。设计了四种方
向的模板，将模板记为T，模板内的元素为0和l，建立的四个模扳为
这四个模板沿着细化后的图像中的细线移动， 依次以每个点为中心与图像做相关运算
c(u,v)为图像中细化曲线上的点，对每一个点可以求得四个不同的相关值， 这个值反映了这个点处的图像与模板的相关程度。其值越大，相关程度越高，
所以当H的相关度最高时，可以认为该点附近的图像的斜率最接近于第k个模板 的方向，因此就可以近似地求出该点的法线。求出了法线以后，沿着法线方向，
以这个点为中心，两边各取6个点，然后利用灰度重心法求出该处的线结构光 ’
图像的中心。

程序实例

clear;clcI = imread('IMG1_1.bmp');if ndims(I) == 3    I = rgb2gray(I);endthresh =20;[r,c] = size(I);for i = 1:r    for j = 1:c        if I(i,j) < thresh            I(i,j) = 0;        end    endend%创建预定义滤波算子gaussFilter =  fspecial('gaussian',[7 7],5);%对图像进行滤波，‘replicate’表示图像大小通过复制外边界的值来扩展I = imfilter(I,gaussFilter,'replicate');%灰度拉伸，增强图像对比I = im2uint8( mat2gray( log(1 + double(I)) ) ); imshow(I);I = double(I)/255;%二值化图像thresh = graythresh(I);I1=im2bw(I,thresh);imshow(I1);%膨胀处理I1=bwmorph(I1,'dilate',10);imshow(I1);%腐蚀se = strel('disk',6);I1=imclose(I1,se);I1 = bwmorph(I1,'thin',Inf);% bwl=bwmorph(I,'skel',Inf);% figure;imshow(I1);% [r,c]=size(I1);% T1=[0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 1 0 0 0 0%     ];% % T2=[0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0;%     1 1 1 1 1 1 1 1 1;%     0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0;%     0 0 0 0 0 0 0 0 0%     ];% % T3=[1 0 0 0 0 0 0 0 0;%     0 1 0 0 0 0 0 0 0;%     0 0 1 0 0 0 0 0 0;%     0 0 0 1 0 0 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 0 0 1 0 0 0;%     0 0 0 0 0 0 1 0 0;%     0 0 0 0 0 0 0 1 0;%     0 0 0 0 0 0 0 0 1%     ];% T4=[0 0 0 0 0 0 0 0 1;%     0 0 0 0 0 0 0 1 0;%     0 0 0 0 0 0 1 0 0;%     0 0 0 0 0 1 0 0 0;%     0 0 0 0 1 0 0 0 0;%     0 0 0 1 0 0 0 0 0;%     0 0 1 0 0 0 0 0 0;%     0 1 0 0 0 0 0 0 0;%     1 0 0 0 0 0 0 0 0%     ];T1=[0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0    ];T2=[0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    1 1 1 1 1 1 1 1 1 1 1 1 1;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0;    0 0 0 0 0 0 0 0 0 0 0 0 0    ];T3=[1 0 0 0 0 0 0 0 0 0 0 0 0;    0 1 0 0 0 0 0 0 0 0 0 0 0;    0 0 1 0 0 0 0 0 0 0 0 0 0;    0 0 0 1 0 0 0 0 0 0 0 0 0;    0 0 0 0 1 0 0 0 0 0 0 0 0;    0 0 0 0 0 1 0 0 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 0 0 1 0 0 0 0 0;    0 0 0 0 0 0 0 0 1 0 0 0 0;    0 0 0 0 0 0 0 0 0 1 0 0 0;    0 0 0 0 0 0 0 0 0 0 1 0 0;    0 0 0 0 0 0 0 0 0 0 0 1 0;    0 0 0 0 0 0 0 0 0 0 0 0 1    ];T4=[0 0 0 0 0 0 0 0 0 0 0 0 1;    0 0 0 0 0 0 0 0 0 0 0 1 0;    0 0 0 0 0 0 0 0 0 0 1 0 0;    0 0 0 0 0 0 0 0 0 1 0 0 0;    0 0 0 0 0 0 0 0 1 0 0 0 0;    0 0 0 0 0 0 0 1 0 0 0 0 0;    0 0 0 0 0 0 1 0 0 0 0 0 0;    0 0 0 0 0 1 0 0 0 0 0 0 0;    0 0 0 0 1 0 0 0 0 0 0 0 0;    0 0 0 1 0 0 0 0 0 0 0 0 0;    0 0 1 0 0 0 0 0 0 0 0 0 0;    0 1 0 0 0 0 0 0 0 0 0 0 0;    1 0 0 0 0 0 0 0 0 0 0 0 0    ];% [H,W] = size(I);% I = [zeros(7,W);I;zeros(7,W)];% I = [zeros(H+14,7),I,zeros(H+14,7)];% [H,W] = size(I1);% I1 = [zeros(7,W);I1;zeros(7,W)];% I1 = [zeros(H+14,7),I1,zeros(H+14,7)];[r,c] = size(I1);H1=zeros(r,c);H2=zeros(r,c);H3=zeros(r,c);H4=zeros(r,c);num1=zeros(11,1);num2=zeros(11,1);centerU = [];centerV = [];center = [];ccc=0;%模板匹配,遍历模板中的点与骨架上的每一个点,如果匹配度最高,必定乘积最大for u=7:r-7    for v=7:c-7        if I1(u,v)            for i=1:13                for j=1:13                H1(u,v)=I1(u-7+i,v-7+j).*T1(i,j)+H1(u,v);                H2(u,v)=I1(u-7+i,v-7+j).*T2(i,j)+H2(u,v);                H3(u,v)=I1(u-7+i,v-7+j).*T3(i,j)+H3(u,v);                H4(u,v)=I1(u-7+i,v-7+j).*T4(i,j)+H4(u,v);                end            end            data=[H1(u,v),H2(u,v),H3(u,v),H4(u,v)];            h=find(data==max( data));            cc=length(h);            if cc>1                h=h(2);            end            if h==1                for k = 1:9                num1(k) = I(u-5+k,v)*(u-5+k);                num2(k) = I(u-5+k,v);                end                centerU = sum(num1)/sum(num2);                centerV = v;                center = [center;[centerV,centerU]];            end           if h==2                for k = 1:9                num1(k) = I(u,v-5+k)*(v-5+k);                num2(k) = I(u,v-5+k);                end                centerU = sum(num1)/sum(num2);                centerV = u;                center = [center;[centerU,centerV]];            end            if h==3                for k = 1:9                num1(k) = I(u+5-k,v-5+k)*(v-5+k);                num2(k) = I(u+5-k,v-5+k);                end                centerU = sum(num1)/sum(num2);                for k = 1:9                num1(k) = I(u+5-k,v-5+k)*(u+5-k);                num2(k) = I(u+5-k,v-5+k);                end                centerV = sum(num1)/sum(num2);                center = [center;[centerU,centerV]];            end             if h==4                for k = 1:9                num1(k) = I(u-5+k,v-5+k)*(v-5+k);                num2(k) = I(u-5+k,v-5+k);                end                centerU = sum(num1)/sum(num2);                for k = 1:9                num1(k) = I(u-5+k,v-5+k)*(u-5+k);                num2(k) = I(u-5+k,v-5+k);                end                centerV = sum(num1)/sum(num2);                center = [center;[centerU,centerV]];            end        end    endend% center=sortrows(center,1);figure;imshow(I)hold onplot(center(:,1),center(:,2),'r.')% hold on% plot(b,a,'b.')