how can i use clarity factor with gaussian kernel in image processing

1 次查看（过去 30 天）
chang hoon oh
chang hoon oh 2021-9-16
评论： Image Analyst 2021-9-23
i apply the gaussian gradient kernel to image for bubble detection
after that i trying to find the clarity of image and remove out of focused bubble. like this article
(Sci-Hub | A fully-automated image processing technique to improve measurement of suspended particles and flocs by removing out-of-focus objects. Computers & Geosciences, 52, 189–198 | 10.1016/j.cageo.2012.08.018)
at the part 3.2.
but i can't find the code for clarity in matlab. so i use the binarize fuction and threshold but it didn't work perfectly so i want to find how i can use
clarity of the image
can anyone help me?
this is my cord
C=imread('30130170023.jpg');
C2=imread('back.jpg');
C1=C2-C;
C3 = imcomplement(C1);
[gx,gy]=gaussgradient(C1,2.0);
[gx1,gy1]=gaussgradient(C3,2.0);
M=abs(gx)+abs(gy);
M1=abs(gx1)+abs(gy1);
MU=M+M1;
Bi=imbinarize(MU,0.3);
BWdfill = imfill(Bi,'holes');
% imshow(MU);
si=imread('sisi.jpg');
si=si(:,:,1:3);
si=im2gray(si);
si=imbinarize(si);
BWdfill=BWdfill-si;
BWdfill=imcrop(BWdfill,[299 0 687 796]);
% imshow(BWdfill);
se = strel('disk',2);
ie = imerode(BWdfill,se);
ie = imerode(ie,se);
ie = imerode(ie,se);
% ie = imerode(ie,se);
% ie = imerode(ie,se);
id =imdilate(ie,se);
id =imdilate(id,se);
id =imdilate(id,se);
% id =imdilate(id,se);
% id =imdilate(id,se);
imshow(id);
[centers,radii] = imfindcircles(id,[3 80],'ObjectPolarity','bright','Sensitivity',0.85);
viscircles(centers, radii,'Color','b');

请先登录，再回答此问题。

采纳的回答

Image Analyst
Image Analyst 2021-9-18
Looks like they define clarity as the max of the gradient in each floc blob:
"The maximum value of Ig associated with each floc i is a
quantitative measure of how close to the plane of focus a given
floc is, and this value is defined here as the ‘‘clarity value’’ a
a(i) = max(Ig(ROI(i))
A threshold......"
  2 个评论
chang hoon oh
chang hoon oh 2021-9-23
yeah i think you are right... but can you give me another advice? i think using gaussian gradient kernel is suit for bubble detection and remove out of focused bubble but sometime it didn't work. apply the threshold to picture some case that erase not out of focused bubble but focused bubble....
Image Analyst
Image Analyst 2021-9-23
It's possible expecially if the bubble had a smooth exterior boundary and is smooth inside. I'd try adjusting parameters like the window side and threshold level.

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Interactive Control and Callbacks 的更多信息

标签

产品


版本

R2021a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by