Field of view correction factor


I’m using imageJ for particle sizing. Particles that touch the edges of the image are removed because their size is unknown. However, because particles that are bigger have a higher chance to touch the edge of the image, they have a higher chance to be removed and thus there is a bias towards smaller particle sizes. I’m using a field of view correction factor (Cfov) for each particle to correct for this. The same as described by Biovoxxel. Using this method, the Cfov of a particle is calculated by dividing the size of the image by the size in which that particle can be measured without being in the “exclusion zone” (see image).

Cfov = (imageHeighthImageWidth)/((imageHeight-BoundingBoxHeight)(imageWidth-BoundingBoxWidth)

However, this method assumes that the particle would always have the same orientation.
Take for example a needle, at 0°C de Cfov would be really low as the bounding box is small. But if the needle would be 45°C it would be really high as the bounding box has a big area. I want to correct for this by using something like the “average bounding box size” if you would take the bounding box from multiple angles. To get a field of view correction factor that takes into account the random orientation of particles. Is there someone who knows the solution to this problem? Or in other words, how would one be able to calculate the chance that a certain particle touches the edge of the image while the center of the particle is still in the image assuming a random orientation of that particle.