Unfortunately, Mercury's disc is so small (about 1/200 of the apparent diameter of the sun) that a telescope is required to observe a transit. Naturally, this must not be done by direct vision as even the slightest glimpse of the sun through a telescope can permanently damage eyesight. Although a dark filter can be used, the best and safest method is projection. In this technique, the image of the sun is directed out of the telescope eyepiece onto a card held some distance away. Not only does this avoid any possibility of eye damage, it also provides an image of sufficient size for easy observation by several people. The card must, of course, be shaded as the image is not very bright: this can best be achieved by pointing the telescope barrel out through heavy curtains. A further advantage of projection is that a small telescope will give an excellent result - my old 60cm refractor was perfectly adequate and much easier to handle than a larger instrument.
In the amateur realm, direct photography of a transit of Mercury is not really practical. Sufficient magnification is difficult to achieve (even with a very large lens) and the fact that the camera will be pointing straight at the full sun makes setting up and focussing very difficult. It is possible to attach a camera directly to the telescope eyepiece, but the same precautions must be taken as with the naked eye and so a sufficiently bright image is hard to achieve. Again, the easiest method is to photograph the projected image: no problems with size or brightness and a close-up effect can be achieved by judicious choice of lens. I allowed the camera to decide its own exposure, which was satisafactory, using 28-70mm and 75-210mm lenses to get the different sizes of image. I also experimented with a x3 close-up adaptor which screwed to the front of the lenses like a filter. This gave very high magnification at the expense of a very small depth of field.
The major downside of photographing the projected image is that geometrical distortion is almost inevitable, as it will be impossible to hold both the telescope and the camera square-on to the card and the image is likely to be slightly off-axis anyway due to movement across the field of view. In addition, variations in the exact lens settings used each time and the distance from the camera to the card make successive images of slightly different size. These distortions make it difficult to compare images and to combine them into a composite, so must be corrected in some way.
Because of the non-linear way some of the distortions are caused, full correction is usually impossible with simple image-processing techniques. Simple variations in size are easy enough to deal with, by re-scaling, as are rotations. I attempted to circularise images which were oval (due to having been taken "on the slant") by rotating until the lengthened dimension was fully either "North-South" or "East-West" then rescaling this dimension to match the other one. This procedure was reasonably successful but did not work so well when the images were "egg-shaped" due to a combination of distortion effects. In these cases, re-scaling just part of the image and then re-combining it with an unchanged remainder gave a result which was visually acceptable even if not entirely scientifically accurate! It was, of course, helpful to know that all of the images of Mercury must lie in a straight line and to use the three large sun-spots as fixed points. This knowledge was a useful guide in determining the degree of rotation and expansion to apply, for example.
After processing, the "full sun" images were finally rotated to place the sun's north pole at the top, with its eastern limb on the left. Other images were placed so as to minimise the amount of blank sun displayed, to save space while still retaining reference marks such as the large sun-spot and the sun's edge. The images displayed here are only a subset of those actually available, as one image of a black dot on a grey background can become much like another after a while!
Back