Methods to determine the size of an object in microns
Mathematically
Pixels on our CCD are 6.45um. The size (in microns) of one pixel in an image will depend on how much the image was magnified before reaching the CCD:
[math]\displaystyle{ \mbox{Pixel size} = \frac{6.45um}{\mbox{total magnification}} }[/math]
This formula gives us:
- 60x : 0.1075 um/pixel or 9.30 pixels/um
- 40x : 0.1613 um/pixel or 6.21 pixels/um
- 20x : 0.3225 um/pixel or 3.10 pixels/um
- 10x : 0.645 um/pixel or 1.55 pixel/um
Important notes:
- We assumed here that the magnifier is at position 1x (not 1.5x). If it is at 1.5x, total magnification must be multiplied by 1.5.
- We assumed here that the bin size is 1x1. If it is, say, 2x2, the size of a pixel will be double.
Calibration slide
Use a calibration slide which has a grid with known line-to-line spacing (9.9um). Align the slide so that the grid lines are parallel/perpendicular to the x and y axes. I borrowed such a slide from Peter Sorger's lab.
- 60x
- 13 squares gave 1204.26 pixels in length. 1204.26 pixels/(13 square * 9.9um/squares) = 9.35 pixels/um
- 14 squares gave 1297.30 pixels in length. 1297.30 pixels/(14 squares * 9.9um/squares) = 9.36 pixels/um
xy-motorized stage
Put a sample on a slide or pad (cells, sphere) or find a grain of dust on the slide or pad. Record the position of the sample. Use IPLab to tell the stage to move a certain distance in um. Determine the distance (in pixels) between the sample's former position and its new position.
To test the calculated um/pixel for a given objective, one could do the following: Get the calibration slide. Take a picture. Knowing how wide (in pixels) our CCD camera is, determine the width (in um) of the CCD image. Move the stage by that distance. Take another picture. Put the two pictures next to one-another. Do the grids line-up?
