Sunday, 27 July 2008

Panoramic and virtual reality photography

I like taking panoramic photos. Some of the header photos on our very slowly developing website (www.thompsonz.net) are early attempts. More recently, I have stumbled into virtual reality photography.

Panoramas

Often with my panoramic photos, I simply take a sequence of photos hand held and without restricting the camera to manual exposure or focus settings. This works fine provided you align the horizon appopriately. This image of Edinburgh comprises three photos taken with my canon EOS300D is an example. The focal length was 34 mm for all of the original photos but the exposure was 1/250 sec at f/13, 1/320 at f/14, and 1/400 at f/14. They were stitched together with Canon's PhotoStitch software. In this case, it is difficult to see the seems between the images.


Original photos:



Stitched result:


Not all attempts are so clean. Here the merge is of two images of the Rimutakas taken from near Featherston in the Wairarapa, New Zealand. On my first attempt, I changed the focal length so the two photos would stitich. But these two photos have the same focal length of 200 but there is enough difference in the exposure (1/500 sec at f/9.0 and 1/800 at f/10) to cause problems in the stitching.

Here are three different attempts. The first with the Canon PhotoStitch, the second with Photoshop Elements 2.0, and the final one with Hugin. Hugin has compensated for the differences in exposure to produce a well blended image.


Original photos:



Canon PhotoStitich image:



Photoshop Elements image:



Hugin image:



VR 360 cylindrical image


With virtual reality photography, you don't simply take a panoramic image in which three or four images are stitched together. Instead you take a set of images that encompass the full spherical view. My reading suggests that most of these are made using a fisheye lens (extremely wide angle). Unfortunately, I don't have such a lens so I am having to experiment with my standard lens. Just to get a full 360 degree horizontal image takes 16 photos so to ensure it aligns you really want a good tripod and a head like Manfrotto's 303SPH.

I decided that I needed to check my locations before taking all the camera gear, tripod, and head. So I took my Canon SX100 IS camera with me and checked the location on one of my bike rides. I then went back as the sun was about to set with all of my gear and set up to take the image.



Initially I viewed it with Canon PhotoStitch Viewer in 360 mode to give the impression of a cylindrical image. The same effect can be achieved by using a Flash or Quicklime if the image is converted to the appropriate format. The following links show the trials in creating a view for the web. Have a look.


Image viewed with Flash viewer
Image viewed with Quicktime viewer

Saturday, 19 July 2008

Seeking employment

To add to the frustrations, I have been approaching employment agencies to find work in the computing industry in New Zealand. Almost a standard response is that I have been out of industry too long and my knowledge of the tool sets is inadequate. I acknowledge that I am lacking in some of the technical details but doesn't the knowledge of the underlying principles and concepts count for anything?

With over 17 years of industry experience from 1975 to 1990 and now 17 years of academic experience teaching students to program and develop systems, you might think that I have something to offer industry but it may appear not.

This isn't a new experience for me. I switched to teaching when I was told by agencies that my mainframe knowledge wouldn't be applicable to programming micro-computers (PCs). You are only good to be employed for the technologies that you know and not for the concepts and principles that you understand and can apply. Having become the guru on AS/400 systems over a three month period when supposedly I was simply a mainframe Cobol hack doesn't seem to count. Now, it seems that for some agents even that extensive Cobol experience is too dated to be able to put me in a position that urgently requires Cobol experience.

All I can say is that the computer industry needs to wake up if it wants to address the shortfall in staff. You can't reject people for the wrong type of experience when you are not recruiting people at the lower end. Yes, the pipeline into Information Systems and Computer Science programmes has shrunk. Students simply haven't been enrolling and now these departments have downsized or are downsizing. New graduates are simply not entering the pipeline. Now, I find at the other end, people may be being discarded from the industry simply because their technical skill sets don't match. That sounds like a recipe for a shrinking workforce.

All I can say is “wake up folks and look at the potential that you are discarding. If you don't wake up then you want have the support that you need for those systems that you want to see put in place.”

Valuing of ideas

This has been a frustrating period as I have looked for employment and endeavoured to complete the writing of my thesis. My supervisor retired this week from the university and as yet, I don't know what the arrangements are for continued supervision. I might be the student but the issue isn't discussed with me and I don't seem to be being notified of progress.

But the frustrations with writing go further than that. It is increasingly looking like I am writing for an unknown target audience who could easily reject my work because their understanding of the issues and approach to research disagree with the premises that I have been working to.

We all know the difficulties that some historical entities have had in getting their ideas accepted. These include the challenge to the earth being flat and to the idea that the sun went round the earth. We are in a much more enlightened age when ideas would be examined fairly based on the supporting evidence. “Yeah right!” Kuhn (1996) talked about how research is guided by paradigms. Well my research crosses over two domains (Computer Science and Education). I am beginning to think this was a dangerous approach. Those in education have little understanding of the computer science issues and the computer scientists lack the educational background. I am currently trying to write from an educational perspective but this is proving difficult because I am not immersed in educational thinking and research paradigms. I write something and it is torn to shreds by my educational based supervisor. I haven't followed the norms of an educational thesis.

But is it a computer science thesis? The answer to that is probably not either. It may not even fit a computer science education perspective. It is beginning to look as though thesis writing is all about knowing the research paradigm in which you sit and ensuring that you stick rigidly to the constraints of that paradigm. Don't attempt to cross the divide or possibly challenge the existing norms too much or you will be rejected regardless of the quality of your work.

Yes, I am assured that my process (data gathering and analysis) is fine. I am just not writing for the target audience. Somewhere I have heard that before. Oh, yes! Secondary school English! Sorry your critique of this play / novel isn't consistent with the normal pattern in English. Grade – Fail. Oh well, let's get on with something that is accepted and leave the critiques to those who follow the accepted norms.

References:

Kuhn, T. S. (1996). The structure of scientific revolutions (3rd ed.). Chicago: University of Chicago.