What is exact meaning of the perimeter?

[redfield]

I trouble about the particle information.
In my application, After processed the blob, and
I reported area and perimeter each indexed particle.
but In the first image, the area is 6, and the perimeter is 8.
What happen?

Before I knew, the perimeter is bigger than the area
But the perimeter is bigger in case of small Particle,
(Image 1)
and the area is bigger in case of big particle.
(Image 4)
So, let me know about How can analyze the perimeter in detail.

I attached image files.

[BruceAmmons]

The area is the number of pixels in the particle. The perimeter is the number of pixel sides on the edge. For example, your first particle is a 2x3 rectangle. The area is 6, which is the number of pixels. The perimeter is 10, which is the number of pixel edges (2+3+2+3) going around the edge.

A particle with smooth edges will have a relatively low perimeter. A particle with a lot of jagged edges will have a relatively large perimeter.

Bruce

[redfield]

Thank you for your answer.
but Please look at a atteched image again.
In the image, you will see the information of the area, perimeter.
so, please check again about the perimeter analysis.

[BruceAmmons]

I appear to be wrong. I'm not sure how it is actually calculated.

Bruce

[ColinAE]

Please look at a atteched image again.
In the image, you will see the information of the area, perimeter.
so, please check again about the perimeter analysis

Indeed, the perimeter calculations that you appear to be seeing are incorrect. The Particle Analysis perimeter algorithm uses sub-pixel interpolation to calculate the perimeter of the particles. When you have very small particles, or very square shapes, the perimeter calculations loose some of their accuracy. For example, I tested on a rectangle that was 97x80 pixels (actual area = 7760, actual perimeter = 354), and the algorithm returned a perimeter of 350, a much smaller percent error than what you are seeing.

With larger, more irregular objects, the perimeter calculation is mo
re accurate with the interpolation algorithms.

Hope this clarifies the behavior you are seeing.

Regards,
Colin C.
Applications Engineering