起初,采取了简单的for循环,手动设定阈值,发现时延严重,代码如下:
vid = videoinput('winvideo',1);
preview(vid);
set(vid,'ReturnedColorSpace','grayscale');
while 1
pause(1);%1s
figure(2);
%imshow(image);
[m,n] = size(image);
for x = 1:m
for y = 1:n
if image(x,y) > 130 %手动设定阈值
image(x,y) = 0;
else image(x,y) = 255;
end
end
end
figure(3);
imshow(image);
end
为了减小时延,采用了matlab内置函数im2bw进行二值化,并用graythresh函数自动获取阈值,设定pause延时为最小25帧(0.04秒),效果大大改善,实时性也很好,代码如下:
vid = videoinput('winvideo',1);
preview(vid);
set(vid,'ReturnedColorSpace','grayscale');
while 1
pause(0.04);%25帧
image = getsnapshot(vid);
thresh = graythresh(image);
image = im2bw(image,thresh);
imshow(image);
end