LabVIEW
calculating volume from an image
Solved!
Hi Omar,
@Omar-Abdelhameed wrote:
The issue I am facing now is when the position of the opject changes, the vi wrongly calculates number of lines!! is there are way to fix that?
Sure: avoid position changes!
More seriously: you could develop an algorithm to crop the image to the interesting part by removing "black borders". This way you only need to analyze very similar images, no matter how large the border area actually is…
Hi Omar,
@Omar-Abdelhameed wrote:
can you please help me applying that! I have never done something similar before.
You never played with GIMP or Photoshop to auto-crop an image?
You never thought about "how did they do that?"?
Building an auto-crop algorithm:
- Define a limit to check for "background" vs. "real image content" and compare the image data against that limit
- In the resulting image: remove all rows/columns that only contain background pixels
- extra step: add a new border of defined width around the "real image content" or center the "real image content" inside a larger background image of fixed size. This will ensure your next analysis steps will always work with images of known/fixed size…
(I will not write the code for you.)
@Omar-Abdelhameed wrote:
I got that solved. please see attached picture. The issue I am facing now is when the position of the opject changes, the vi wrongly calculates number of lines!! is there are way to fix that?
I appreciate your help and support. Thank you.
I am pretty sure this code is bugged if you look closely (if the input is a rectangle, the output should not be be square!):
To see how to rotate or flip a 2D array, look here.
If you take a horizontal and vertical profile of the brightness, you can tell from the pattern what the correct orientation would be and adjust the code accordingly. You can also tell where the black borders are and only use the inner subset for further processing.
