Calculate the perimeter of object in binary image how!!!!!!!

[mangood]

Hi 

 I try to use labview to find area and perimeter of object in binary using vision function

for the area the labview find it in pixel]

but for   perimeter I have problem I read from ni.com vision manual this words

http://zone.ni.com/reference/en-XX/help/372916M-01/nivisionconcepts/particle_measurements/

"""Length of a boundary of a region. Because the boundary of a binary image is comprised of discrete pixels, NI Vision subsamples the boundary points to approximate a smoother, more accurate perimeter. Boundary points are the pixel corners that form the boundary of the particle. Refer to the introduction for an illustration of pixel corners."""

what is the problem I really did not understand what he mean::"""Because the boundary of a binary image is comprised of discrete pixels""

can any one explain to us the case please and who will lv find the  perimeter

best regards

[Hornless.Rhino]

---

@mangood wrote:

what is the problem I really did not understand what he mean::"""Because the boundary of a binary image is comprised of discrete pixels""

can any one explain to us the case please and who will lv find the  perimeter

best regards

---

Take the following image on the left. You can see it is quite jagged.
Since pixels are square/rectangular you cannot have a perfectly straight edge on a line that is not horizontal or vertical.

But looking at how the edge of the object actually is, the jaggedness can be "smoothed out" and hence help determine what the perimiter should be were it perfectly smooth.

[mangood]

---

@Hornless.Rhino wrote:

---

@mangood wrote:

what is the problem I really did not understand what he mean::"""Because the boundary of a binary image is comprised of discrete pixels""

can any one explain to us the case please and who will lv find the  perimeter

best regards

---

Take the following image on the left. You can see it is quite jagged.
Since pixels are square/rectangular you cannot have a perfectly straight edge on a line that is not horizontal or vertical.

But looking at how the edge of the object actually is, the jaggedness can be "smoothed out" and hence help determine what the perimiter should be were it perfectly smooth.

---

thank you 

do you mean that lv do a low pass filter for image before he caculate the  perimeter?????

finally which connectivity is best 4 or 8 for finding the perimeter??? or which algorthem used by lv to find the  perimeter???

best regards

[Hornless.Rhino]

---

@mangood wrote:

---

---

thank you 

do you mean that lv do a low pass filter for image before he caculate the  perimeter?????

finally which connectivity is best 4 or 8 for finding the perimeter??? or which algorthem used by lv to find the  perimeter???

best regards

---

have a look at http://forums.ni.com/t5/Machine-Vision/Vision-perimeter/m-p/1787588#M33709