Length of a thin curved object in an image

Question

Audrey 2018-1-8

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376195-length-of-a-thin-curved-object-in-an-image

评论： Matt J 2018-1-10

采纳的回答： Matt J

testimagecurved.png

I'd like to use image analysis to calculate the length of a thread.

Because the thread is not entirely straight, I can't just use regionprops, so I think the best option is to use the function bwdistgeodesic

In the bwdistgeodesic function, what is it expecting for the variable "mask"? How could I use image analysis to generate said mask? I've looked into using regionprops 'Extrema' property, but since I'm not sure what the "mask" variable is in bwdistgeodesic, I'm not sure how I would use the x,y dimensions from 'Extrema' in a mask.

I've attached the example image I'm using - with a "thread" and a one inch coin for size reference.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Matt J 2018-1-9

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376195-length-of-a-thin-curved-object-in-an-image#answer_299248

编辑：Matt J 2018-1-9

在 MATLAB Online 中打开

Because the thread is not entirely straight, I can't just use regionprops,

Why not? Couldn't you just do

   S=regionprops(  bwareafilt(BW,1,'smallest') ,'Perimeter');
    length=S.Perimeter/2

4 个评论
显示 2更早的评论隐藏 2更早的评论

Audrey 2018-1-9

Actually, the new issue I'm having is that the perimeter edge is very jagged, and it's giving me a longer value than expected. I wonder if this can be improved by reducing the size of the image and using sgolayfilt. I'm not sure if that can be used on the image itself, or if I have to draw the perimeter line, and then apply the filter to that.

Matt J 2018-1-9

I think you could just extract the points using regionprops(...,'PixelList') and then apply filtering.

请先登录，再进行评论。

Answer 2

Image Analyst 2018-1-9

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376195-length-of-a-thin-curved-object-in-an-image#answer_299352

在 MATLAB Online 中打开

Assuming your line is just a one pixel thick line, you could just compute the area with bwarea() which weights the area depending on the shape of the boundary.

binaryImage = bwareafilt(binaryImage, 1, 'smallest'); % Extract smallest area.
% Now only the line is remaining.
lineLength = bwarea(binaryImage); % Get area, which is it's length.

2 个评论
显示无隐藏无

Audrey 2018-1-10

Hi IA,

The future goal is to be able to do this with a real image, and when I tried this (with image below), the binary image produced a line that was many pixels wide, and a bit jagged. I'm working on using filtering to smooth this, but if you have other ideas let me know!

here's that picture: