The Art and the Science of Creating Panoramic Views

Where is that View?  To get into this hobby, I had to go away from the place where I grew up. Vorarlberg, where I live now again, is the westernmost province of Austria, known as an alpine country. Most of our teachers liked outdoor activities, so the regular school hikes at the beginning and at the end of the school years led us somewhere up into the mountains. And I hated it at that time. Not being used to it and not being well equipped meant nothing else than blisters on the feet and aching muscles the next day.  I would have much more preferred to spend these days in the valley at a pool ...

When I returned to Vorarlberg in 1998 to work for OMICRON,  Idiscoverd the beauty of my home place. I had been away for 17 years while attending university and working in Germany. Now being back, I started hiking in our mountains by myself. Accidentally, I also obtained a homecoming gift at that time, the book "Vorarlberg Panorama 360°? fromWaltraud and Walter Vonbank. Walter Vonbank worked for the local electric power generation company as a full time photographer. He used a Swiss made Seitz Roundshot camera to create impressive panoramic images already in the 1980's. The camera had a vertical slot lens and was rotating on a tripod. It was expensive, heavy to carry, and sophisticated to operate. All of this seemed so exotic that I never thought it would ever become a topic for me.             

When preparing for my hikes, I often make an internet search to get hints on hiking trails.  Remarkably often, the search results offered links to a website with panoramic pictures (or in short "panos"). These panos were made from series of images shot with digital cameras.  It is much easier to do this now, since no expensive special equipment was needed any more. Some of the panos were carefully labelled and had other interesting explanations attached.

I realized that I had been at some of these observation points and I started browsing through my digital photo archive. And indeed, I found a few photos that overlapped and could be stitched together to make up a little pano. However, the panos would have been quite nice if I had shot some more pictures to expand the view. 
Thus, from this moment on I tried to make series of photos to create panos out of them. I also created an account at the pano website and started posting my own panos (http://www.alpen-panoramen.de/portfolio.php?id=9317075).

The Pano Community: While browsing the panoramic website, I quickly identified the hard core of the community and the champions making the best panos.  I started communicating with some of them to ask for hints.  First on the trails to get to the observation points, then about technical details of making better panos. I did not have much experience in photography and when shooting a pano with a wide viewing angle you typically have largely varying light conditions depending on the viewing direction.  I had troubles dealing with this and my series often contained bad photos that spoilt the whole pano. I got hints on how to adjust white balance and exposure which helped me very much.

As supportive the community is, it is also demanding.  If your panos are noticed at all (and that is what you want), they get rated by the others.  You easily get low ratings for bad photographic quality or stitching errors. And it is expected that at least the major summits visible on the pano are labelled with name and elevation. Failing to do so (being lazy) can lead to negative comments and degraded ratings. When a pano remains with too bad ratings for a while, it will be automatically removed. This has happened to me in a few cases.

My own Level of Sophistication: Most modern digital cameras have some kind of pano function that allows easy creation of panos by just moving the camera over the field of the view. But in most cases, the resulting pano is limited to HD resolution (1080 pixels in height) and lacks further control of quality.  I definitely wanted to go a step further with stitching individual images of higher resolution and optimized exposure.

On the other hand, the champions use even more sophisticated techniques and gear. Some of them carry many kilos of equipment up the mountains. They use tripods with levels for exact horizontal alignment, and panorama plates that allow pivoting of the camera in exact angle increments between two images.  Some of them even have nodal adapters, but I doubt that they use it for mountain panos, because it has no benefits when making photos of far distant objects.

I refuse to go so far, so I shoot my images "free-hand" and just hope that the series of pictures works out for a reasonable pano. The challenge with free-hand shooting is not to "lose the horizon". To illustrate this problem, I created a sample pano (Figure 8).
It was shot with a tripod, the images on the left and right level perfectly, but the one in the middle was heavily "dropped" with the purpose to reveal the effect. The yield in height is dramatically reduced to a degree where essential parts of the pano become cropped, thus destroying the whole pano. By the time I have developed some skills in keeping the horizon, so the yield in height is quite good in most of my panos. My new camera has a viewfinder with an integrated level gauge that helps very much in this regard.

Most power users from the pano community invest in software and use commercial image processing programs for stitching and post processing the panos.  I use an open source stitching software. It is not super user friendly and has quite a learning curve, but its features are very good. 

Some champions suggest to use the Raw image format of the camera to get even better results. It is said that correct expose is much less crucial when using Raw format. But the overhead is considerable, so I just go with the JPEG images.

Locations, Directions and Mapping:  I use a GPS navigation device to find my way and to track the paths I go, also to get the position data of the observation point where I take my panoramic pictures. When posting a pano, the specification of the coordinates of the observation point is important.
Beyond the labeling, providing information about the viewing direction adds to the quality of a pano.  When I started with this, I got into a discussion with a power user from Switzerland who provides such information on most of his panos. But our viewing angles were different. Not much, but noticeable and consistent in magnitude. So it was not an uncertainty, or a calculation error, but something systematic.

This got me deeper into the mapping topic than I expected. I traced this issue down and found what professional cartographers could have told me right away.  The reason is the Mercator map projection, which is the underlying principle of most topographical maps in use today. The projection is on a rectangular grid, but only at the reference meridian of the maps the "grid north" aligns exactly with the true north direction. At positions off the reference meridian, a difference occurs which is called the "meridian convergence".

My peer was using the Swiss map system, and I was using the Austrian map. As the maps have different reference meridians, they are slightly tilted against each other. This was the difference we observed, it could also be obtained by adding up the meridian convergences at our observation point in the two map systems.

The vendors of the GPS navigation devices also offer topographical maps with hiking trails. But these maps are not cheap and are soon outdated, since even hiking trails change over time. On the other hand, the local land surveying offices maintain maps for all kinds of purposes. You can view up-to-date hiking maps on their websites. When preparing a hike, I make screenshots of these maps to get map tiles of the area where I am going to hike. Then I create custom maps for the use on my navigation device.

Therefore, the map tiles have to be connected to the corresponding geographical location, which is called georeferencing. The recommended method for georeferencing is using Google Earth, which works well if sufficient details can be seen on the satellite images. I noticed that I had to rotate the tiles slightly to make them align with the satellite image. I determined the suited rotation by trial and error. As I learned later when looking into the direction issue mentioned above, this rotation angle must be the meridian convergence and a check of the values confirmed that they matched quite well.

But when the satellite images were taken in winter, the snow hides most details and the images are uniformly white. The recommended method for georeferencing fails. Therefore I developed a method to georeference such a map tile by knowing the coordinates of two points on the tile. I implemented this method as a Python program, and again understanding the Mercator projection was crucial for this.

Thus, my hobby evolved into several related activities, most of which have to do with wonderful technical gadgets like digital cameras, or GPS devices that are perfect toys for engineers. It is interesting, because there is always something new to learn.
And it is not just the outdoor activity for the days with good weather, there is also the preparation and post processing, that keeps me busy on rainy days at home.  

Biography:

Fred Steinhauser was born in Austria. He studied Electrical Engineering at the Vienna University of Technology, where he obtained his diploma in 1986 and received a Dr. of Technical Sciences in 1991. In 1998 he joined OMICRON, where he worked on several aspects of testing power system protection. Since 2000 he worked as a product manager with a focus on substation communication issues. Since 2014 he is responsible for Power Utility Communication business of OMICRON. Fred is a member of WG10 in the TC57 of the IEC and contributes the standard IEC 61850. He is one of the main authors of the UCA Implementation Guideline for Sampled Values (9-2LE). He is also a member of SC B5 of CIGRÉ and contributed to the synchrophasor standards IEEE C37.118.1 and IEEE C37.118.2.

Protecting your electrical assets? today and tomorrow