Never seen aerial virtual reality images before? Here's how they are made:
They are a combination of aerial photography and Virtual Reality photography. Aerial Photography is what this site is about, VR is added to give the viewer an even more intense experience.
VR uses extremely wide angle lenses to capture as much of what is around the camera as possible, typically 3 or 4 images are taken with a fisheye lens at the four points of the compass, then optionally a shot pointing straight up, and one pointing down. These are then merged using special software into an equirectangular image which covers 360° from right to left, and 180° from top to bottom.
With a kite the platform is rarely stable enough to perform this work successfully, as the images must be taken from exactly the same point so there are no visible errors in the finished product. However with aerial shots there is rarely anything above the horizon, which allows a simplified approach to be used, as pioneered by Philippe Hurbain.
Until 2011 the camera
was flown in a special rig with a single lens pointing straight down (e.g. RIG05,
RIG06, RIG07). As the field of view is 183-8°
then the resulting image includes all the subject and the entire horizon as in the example on the left.
With nothing of interest "above" the horizon, then Philippe's technique is workable, however the resulting image, based on one lens, has limited resolution.
The revised equipment used from 2012 uses 4 syncronised cameras covering
the entire area below the horizon (e.g. RIG14,
and RIG15). The field of view is still greater than 180°
so the resulting image still includes everything of interest, but with greater resolution.
In 2018 RIG16 was introduced with better optics and syncronisation, producing crisper and easier to stitch images, using the same mechanical rig.
In both methods multiple shots are taken during the flight, and those without the horizon visible as a complete circle are discarded.
After the aerial shot was
taken, and the camera recovered, I entered the castle and climbed to the top of
the tower on the right of the above picture. By pointing the camera straight up
I get a second shot, shown right.
For effect I flew the kite from this location, although at this
point the camera is no longer on the kite. From this high vantage point the
horizon is clear, with the exception of one small part of the tower roof.
After the shoot special
software is used to produce a equirectangular version of the image from the fisheye
original.
The circular horizon has become a straight line, and the point
immediately below the camera has become a line along the bottom of the picture,
in the same way as the South Pole on a Mercator Projection map of the world.
The top image is produced
from the older rigs with fisheye lenses, below the result from the newer, higher resolution 4 camera rig.
The 4 camera version produces the same result but with a scalloped sky,
the key point is that the lower hemisphere is the same, we will move onto the sky next...
The same transformation is done with the sky,
then these two images are combined in a package like Paint Shop Pro or Photo
Shop, using mask layers to blend the sky into the shot.
At this point the remaining roof is edited out, scallops removed, the sky is colour matched to the ground, and the sun is placed so the shadows are correct.
Sometimes a "stock" sky is used if it is not possible to get a good view from the location.
The final image is carefully
edited along the horizon, to make the transition from sky to ground as natural
as possible.
Note that it isn't important that the kite is clipped at the edge of the image, there is no edge, they are the same point!
Finally this image is used with HTML5 or Flash Player (although in the
past I have used Shockwave, Java, and Quicktime) to
produce the viewable version that the web-viewer sees.