The Transit of Venus on June 5, 2012, observed with a pinhole camera
G. Gwinner, Winnipeg, Manitoba (gwinner@physics.umanitoba.ca)

Getting there
My 8" Schmidt Cassegrain is still not unpacked after 9 years in Manitoba, and I didn't get aluminized mylar foil for filters. Planned to content myself with watching the transit on the web with the kids. But then the weather was just too perfect and I started looking around for interesting ideas. What about a pinhole camera? Would the resolution be good enough to see Venus (57 arcsec disk)?
The only reference I found was Hugh Hunt's "pinhole mirror" observation of the 2004 transit at Cambridge
Due to diffraction kicking in at small hole diameters, a long projection distance is required - too long for a cardboard tube style camera. Hence the idea of using my hallway as the projection chamber.

The setup
The only window in the house not obstructed by trees faces north
Use a "heliostat" mirror outside on the garage roof to bring the light into the hallway

The pinhole was attached to the window, a Thorlabs iris

A first projection of the sun

The transit of Venus
The transit starts, as can be seen on the live stream from the University of Manitoba observatory, and right afterwards, we see it with our pinhole camera (at south-south-east)

Turns out I was too greedy with the pinhole, made it too small (approx. 1 mm), and diffraction was too strong, opening it to 2 - 3 mm, gave a much better picture

An aside on diffraction: there is a very nice plot on Wikipedia
Click to enlarge

This shows nicely that 1 mm aperture only gives you on the order of 100 arc sec resolution.
The actual image was a lot better than the photos. Turns out the bottleneck was the autofocus of the point-and-shoot camera!

Comparing our projection (left) to the UManitoba live stream on the iPad (right). Again, the picture does not do the projection justice.

A last look, and off to story reading with the kids. They will be 113 and 111 when the next transit occurs in 2117. Time to focus on transits of mercury, which are more frequent but will most likely not be visible with a pinhole camera.

Some more thoughts on a pinhole observation of Mercury
As the chart above shows, we would need about a 1 cm aperture to reach 10 arc sec resolution in terms of diffraction, and mercury is about 13 arc sec in diameter during a transit.
The sun is 60*30 = 1800 arc sec in diameter, and Mercury around 10 arc sec, so we need to resolve the sun's disk to 1 in 180 in order to spot Mercury. If the pinhole is 1 cm in diameter to satisfy the demands of diffraction, the sun's projection image would have to be 180 cm in diameter. This would be achieved with a projection distance of 225 meters! Quite long, but not impossible. Folding the path of the light with mirrors is not possible, as the light cone is soon very large and standard wall mirrors will degrade the image quality too much. But it should be easy to project in the outdoors and only do the last few meters inside a dark room (like Hugh Hunt did). It would be quite necessary to automatically move the mirror to keep the image of the sun fixed in one spot.
However, the image brightness would be much reduced. The distance is 30x larger and the aperture is maybe 5x larger, so the surface brightness would be down by a factor (30/5)^2 = 36. That's quite a lot, on the other hand, the image was quite bright and the hallway had not to be darkened seriously. There is still some room there.