Handheld photomicrography and reticle

kw: photomicrography, photographs, cameras

Having only a laser printer at my disposal, I've been trying to make object-plane reticles to calibrate the magnification of photomicrographs. This first, poor item is 5x5mm, with the light and dark bars intended to be ¼mm (250µ) wide. However, uniformity is awful:

Next I tried a small square containing other squares. My initial attempts were aimed at having the smaller squares be 0.1mm (100µ), but after various tests, I settled on the square being 2.5x2.5mm, with the small squares being 250x250µ. If you click on this image, the larger image you'll see has a magnification of about 52X. This one is closer to 15X. The target insect is a midge about 1.2mm long in body, with legs approaching 2mm long:

This last image is a 1x1mm clip from a more highly magnified original, centered on the midge's head. At this magnification, 1mm = 480 pixels, so if you click on the image below, the larger version, being 480 pixels across, will occupy 4.8 inches on your screen, or a magnification of 122X:
With this microscope, at least, handheld photomicrography works pretty well. I am considering a jig that will screw into the tripod mount and attach to the microscope tube, so I can get more repeatably aimed results, and even make short videos (like of the critters in my birdbath). But for most purposes which I use this microscope, this method is sufficient.

Calibrating for low power microscopy

kw: photographs, photomicrography

NOTE: This post has large (400px) images side-by-side, so is best viewed full screen.

I took advantage of the holiday sales to purchase a small digital camera suitable for taking photomicrographs by pointing it into a microscope eyepiece. It is a technique I've used with film cameras in the past, starting fifty years ago when I used my father's Argus 35mm camera with my little "kit" microscope. My current (now "main") digital camera, a Nikon D40, has a lens much too large for this use. The eye relief of the eyepieces of my microscopes ranges from 9mm to 14mm. A camera's lens has to be short enough that the eyepiece's focus point can be put in or near the center of the iris diaphragm, or the image will suffer vignetting. A Canon SD1200, with its 6.2-18.6mm zoom lens, seems well suited to this use, and the price was right. I had already learned (reported here) that my son's SD1100 was useful for this purpose.


Most cameras these days have a Macro setting, and the Canon's is pretty good. These photos show the practical range available. The first is the whole image of a portion of a post card and stamp. Reproduced this size, its magnification is 2x on a 100dpi monitor, or 2.3x on a 86dpi monitor. However, there are a lot of pixels in the original image. I have the camera set to 6Mpx (2816x2112) to match my DSLR's resolution. It also gives it better low-light performance. The second image (on the right if there is enough room) is a 400x400 pixel crop, and has a magnification of about 13x on a 100dpi monitor. A 6x4 inch print will have a magnification of 3x. Cropping out a 1200x800 section, a 200dpi print's magnification will be about 7x. That's a good working range for many purposes.

I proceeded to calibrate the camera with my stereo low-power microscope. Its visual magnifications are 7x, 15x, and 20x. I have a set of eyepieces that take the range to 30x, but the eye relief is smaller, making it harder to line up the camera. The third and fourth images are of the millimeter scale on a wooden ruler, using the 20x setting on the microscope and the camera's lens zoomed all the way "long". Direct scaling of the digital image yields a factor of 660 px/mm. The other image of the scale is an 800x800 crop, further reduced by Blogger to 400x400 (click to see the larger one).

Maximum displayed magnification is 168x, and maximum printed magnification is half that. Displaying the entire image at 400x300 pixel size has a magnification of 24x. Thus I have a set of factors to use if I need to report exact magnifications. Of course, as with a zoom lens, actual focal length is rarely reported, and for many microscopy purposes, reporting the original size of an object is often sufficient.

The fifth and sixth images show part of the butterfly on the stamp. First, the whole image at is shown at 24x, then a 400x400 crop that shows the halftone color dots making up the image, at 168x. This is the butterfly's eye.

Some part of the blurriness of the latter photo may be due to the camera optics, but I think most is the fuzzy edges of the dots themselves. I find it interesting that, as the ink dried on the shiny stamp paper, it migrated to the edge of each dot, forming a ring. I suspect it isn't really supposed to do that, and that the stamp would look better if the dots were solid.

I grabbed another test subject I had on hand: a bunch of insects cleaned out of a light fixture and kept in a plastic vial. Spreading them out with a needle, I first separated a small weevil. Using the 15x setting and the camera zoomed all the way, I got this image (#7), shown here at about 17x (the weevil's length is 4mm, exclusive of antennae).

The last image is more of a survey, meant to show the variety of critters the light had attracted, centered on a small wasp. The image's magnification is 8x. While that is in the range of the macro lens, I can crop a section of this photo to obtain 56x if needed, or 28x for a print.

All these photos were handheld. I have a small tripod, but it will take some fiddling to produce a setup in which the tripod-plus-camera can be quickly put into the right position. That will be more critical with my other, high-powered microscope, which doesn't let through nearly as much light. It may be some time before I am able to calibrate that setup.

Picturing the third dimension

kw: photography, photomicrography, new technologies

In 2000, Noel T. Goldsmith wrote "Deep Focus; a digital image processing technique to produce improved focal depth in light microscopy" (Image Anal Stereol 2000; v19; pp163-167 and this PDF). The article outlines and demonstrates how software can detect the most in-focus parts of multiple photographs of the same subject and stack them together to produce one image that is in focus throughout. It is the 'through-focus' analog of panoroma stitching.

It only took a year or two for the first commercial products to appear, and by 2003 there were several. One that seems popular is Helicon Focus. The three images that follow are from a review of Helicon Focus at Digital Photo Pro. They suffice to show the principles.

This is the first of ten images in a "stack" taken over the wide focal range of this sunflower field.

This image, the tenth in the stack, is close to what you'd get by focusing near the far end of the field, and letting the closest flowers go out of focus.

With the software, the whole field appears in focus. The original paper by Goldsmith was applied to microscopy, and that is still its most useful arena, for depth-of-field problems are the greatest there.


This image, from Concept 2 Innovation, could not be taken any other way. It is actually a stitched mosaic (panorama) of several through-focus stacks processed by Helicon Focus. Of course, this is quite a bit smaller than the research photograph, which shows great detail over and throughout the entire fly. It think of such a photo as a kind of gigapixel image with three dimensions. Producing two such images at an angular difference of 10° will yield a stereo pair of fully focused images. Heaven for a microscopist!

My micrographic roots

kw: microscopy, photomicrography

I entered science fairs from sixth grade right through high school. For most of them I displayed my gradually increasing skills in photomicrography. At that time (1958-1964), I used a rangefinder 35mm camera on a tripod, adjusted to shoot through the eyepiece. The lens was fixed focus (zoom was extremely rare and costly), 38mm. I had to turn the camera away to set up each new shot, then reset it. The camera's focus would be set to infinity. I first learned lighting and exposure, then developed my skills in microtomy and specimen preparation. My junior year of high school, the science teacher actually paid me to take photos of many of his standard slides so he could show them to the students without using the cranky projection microscope the school had. The pictures I made in those days all looked similar to this first image, made with a point-and-shoot camera, handheld (!).

Canon SD1100 IS, f/2.8, 1/125 sec., 1x zoom. Microscope 20x.

I've wanted to get back involved with photomicrography for some time. When I had a working 35mm Yashica SLR, I had a T-mount adapter for it, and did some work with that setup. My digital Nikon SLR will not work with a T-mount adapter. I tried shooting through the lens, but the vignetting is too severe; there is no way to get the entrance pupil (diaphragm) of the camera lens anywhere close to the exit pupil of any of my microscopes.

My son took off for a couple days at school (he's taking a 2-day-per week class) and left his camera behind. It is a point-and-shoot, specification given above, with its 3x zoom lens working in the range 6.2-18.6mm. The focal distance of the 10x eyepieces on the inspection microscope is 25mm, and the eye relief (distance from lens housing to exit pupil) is 20mm, so this seemed ideal. It is!

The specimen is a rice beetle (more on this below). Directly scaled, its length is 2.7 mm. On my 100 dpi screen the image above has the critter's length as 27mm, so this is a 10x view. Click on the image to get a larger one that is quite close to 20x. The original image is 8Mpx, 3264x2448. I reduced it by 4x to 816x612 (the image you see by clicking), and this view is 400x300, set by the Blogger software.

This (click to see full size) is an 816x612 crop from the image above, to show the detail in an unreduced image. This is about 40x, and the full size is about 80x, on a 100 dpi screen.

There is reasonable depth of field. The ornamentation on the beetle's back (the elytra) is beautiful. Only one claw is in focus. The eye appears as a dark bulb, no detail visible.

This beetle is one of a few dozen that showed up when we boiled some old rice. My wife called me to the kitchen to ask what the little black specks swirling in the pan were. I scooped a few out with a spoon and went to the microscope. She was not too happy that they were beetles!

The beetles were floating, and the rice sank, so we swirled the water and scooped them out, several times, until we could find no more. They have sat under my microscope for a couple of months.

A nice thing about modern cameras is that so many now come with good zoom lens optics. The next two images show what is possible with this particular model.

Canon SD1100 IS, f/4.9, 1/79 sec., 3x zoom. Microscope 20x.

This image is very close to 30x. Click on it to see the 60x original. The Ramsden circle (image of the field-limiting diaphragm in the microscope eyepiece) has been zoomed out beyond the image area. Comparing this with the 40x image above, it is apparent that the depth of field here is a bit less. The elytrum ornamentation is still quite clear, but some details around the body are not as well delineated.

By the way, all of these have been color-adjusted for a white background. The automatic color balancing in both Canon and Nikon cameras does a poor job of compensating for incandescent light, and produces yellow-orange images. I know how to force better color balancing in my Nikon SLR, but I'll have to do some reading in the manual to find if this Canon can be "told" to compensate for incandescent light.

This (click for full size) is an 816x612 crop from the image above, showing details on the head. The eye lenses are visible, though not as distinct as I'd like. I suspect, using a tripod or other mounting, that I can force a lower ISO (It was using ISO 120) and slower shutter for less color jitter in the image.

The effective magnification of this image is 120x, and of the full size crop, 240x. That's pretty good for beginning with a 20x setting on an inspection microscope. For making prints, the effective magnification depends on the sort of cropping I do before printing. The kiosk printers at the local drug store print at 200 dpi. If I prepare 1200x800 images for printing at 6x4 inches, I'll get maximum resolution. Starting at the top, then, I would take the 3264x2448 original, which produces about 20x at 100dpi, crop the top and bottom to get 3264x2176 (for a 3:2 ratio) and reduce it to 1200x800 to get a 10x image on the paper print.

Cropping out a 1200x800 piece is the same as a 2.7x magnification, so I'd have a 27x print. The 3x zoom images would then print at 30x and 80x.

Now that I have seen what is possible, I'll investigate the optics of PnS cameras similar to this Canon, and get one for which I can make some kind of mounting (they have a 1/4-inch threaded tripod mount). Also, when pond water is the subject, there's something these cameras can do that 35mm can't: video!