[Digital image processing matlab] (boundary tracking - raster scanning method)
Call Matlab's built-in image test, and the raster scanning method realizes boundary tracking.
function coor = edge_search(I)
%边缘跟踪
%输入：灰度图矩阵，用Matlab自带图像 rice.png 进行测试
%输出：米粒边缘坐标（顺时针）
%调用------------------------------------
%I=imread('rice.png');edge_search(I);
%----------------------------------------
%转化为标记矩阵
%L=bwlabel(BW,n) 返回一个和BW大小相同的L矩阵，包含了标记了BW中每个联通区域的类别标签，n的值为4/8，默认为8（连通寻找）
I = I >130; %得到logical型矩阵
L = bwlabel(I);
%标记连通图像
[row,col] = size(L);
connected = zeros(row,col);
labeled = [];
coor = {
};
curr_coor = [];
%8邻域坐标
offsets = [0 -1; -1 -1; -1 0; -1 1; 0 1; 1 1; 1 0; 1 -1; 0 -1];
flag = true;
for i = 1:row
for j = 1:col
%判断L(i,j)为前景 且 未被标记 且 该区域未被标记
%[ismember(a,b),返回与a大小相同的逻辑数组如果a中元素属于b,a中相应位置返回1，否则返回0]
if L(i,j) ~= 0 && connected(i,j) == 0 && ~ismember(L(i,j),labeled)
%初始方向
dir = 2;
%保存当前区域标签
labeled = [labeled L(i,j)];
pix = [i,j];
curr_coor = [curr_coor; i j];
%坐标入队
flag = true;
while flag
flag = false;
%循环到每圈初始时将方向初始化
if pix(1,1) == i && pix(1,2)== j
dir = 2;
end
%邻域搜索
for k = dir:dir+7
tmp = mod(k,8);
if tmp == 0
tmp = 8;
end
n_pix = bsxfun(@plus,pix, offsets(tmp,:));
%每圈终止
if pix(1,1) == i && pix(1,2)== j && length(curr_coor) > 2 && n_pix(1,1) == curr_coor(2,1) && n_pix(1,2) == curr_coor(2,2)
flag = false;
break;
end
%判断坐标在范围内
if n_pix(1) >= 1 && n_pix(1) <= row && n_pix(2) >= 1 && n_pix(2) <= col
if L(n_pix(1),n_pix(2)) ~= 0
%标记（画图
connected(n_pix(1),n_pix(2)) = 255;
%保存坐标值
curr_coor = [curr_coor; n_pix(1) n_pix(2)];
%更新坐标
pix = [n_pix(1),n_pix(2)];
flag = true;
%更新反方向的下一方向
dir = mod(tmp + 5,8);
if dir == 0
dir = 8;
end
break;
end
end
end
end
%保存一圈坐标
coor = [coor; curr_coor];
curr_coor = [];
end
end
end
%输出显示
subplot(1,2,1);imshow(I);
title('原图像');
subplot(1,2,2);imshow(connected,[]);
title('边界跟踪(光栅扫描法)');
end
test:
