How to find arc length, angle from the image?

Dear all,

I want to find the curve length and curve orientaiton from the image. I performed the image segmentation and convert it to biary image (attached is the iamge). How can I measure the angular variation of the curves and the curve angle.

Is there a better way to perform segmentation of the image to see the more sharp and clear curvatures?

how can it be quantitatively/statistically characterise?

Thanks

Answer 1

[Bhanu Prakash]

Hi Abhishek,

As per my understanding, you have obtained the segmented binary image and want to measure the parameters such as curve length, curve angle and angular variation.

1. To extract the curves from the segmented binary image, you can use the ‘bwboundaries’ function in MATLAB, which provides you with a set of (x,y) coordinates of each curve in the image. For more information on the ‘bwboundaries’ function, refer to the following documentation: https://in.mathworks.com/help/images/ref/bwboundaries.html
2. To measure the curve length, you can use the ‘Arc Length formula’ to calculate the Euclidean distance between consecutive points and summing up the distances to get the total length of the curve.
3. To measure the angular variation of the curves, you can compute the tangent direction on each point on the curve by using the ‘gradient’ function in MATLAB. Please refer to the following documentation for more information: https://in.mathworks.com/help/matlab/ref/gradient.html
4. To improve the characteristics of the segmented image, you can follow the advance segmentation techniques such as Active Contour and Graph Cut method.

For more information on the afore-mentioned techniques, refer to the following documentation:

1. For Active Contour segmentation: https://in.mathworks.com/help/images/use-active-contours-to-refine-the-segmentation.html
2. For Graph Cut Segmentation: https://in.mathworks.com/help/images/segment-image-using-graph-cut.html
3. For more Image Segmentation techniques, please go through the following documentation: https://in.mathworks.com/help/images/image-segmentation.html

Answer 2

[Image Analyst]

See my attached fiber orientation demo.

[Abhishek Saini]

Thanks for sharing the demo. It’s interesting to look those images. I can see some voids there. Is it also possible to automatically capture the dimensions of those voids and plot the distribution?

[Image Analyst]

Sure, you just need to define what is a void.