On 13 Jun 2003 at 10:16, Jon Pumplin wrote: > Instead of trying to instrument the paddle, how about just painting a > pattern on it and taking a movie in stereo? Easy in principal, tricky in practice -perfectly doable. I once spent an afternoon with a film specialist in California next door to the deep sea testing tank where, years later, parts of Titanic were filmed (This was back in the days when I worked for a naval architecture firm specializing in ice mechanics - ice breakers, arctic oil exploration platforms etc). The lessons learned didn't greatly advance my own activities on filming tank tests ($$$ vs time vs willingness of clients to pay - spending $100,000 on a series of tank tests and then _not_ gathering good data seemed to most clients to be a good investment... Sigh - the lack of good data usually lead to further tests. Since this made my boss richer, he thought it was just fine.). The rather crazy California guy had created model oil rigs (semi-sub) with little grain-of-wheat bulbs (LEDs weren't well known by him) stategically placed on the model and then filmed the launching sequence in the tank using special (military surplus) high-speed registration film cameras. He then played the film back through a registration projector onto the back of a screen and _manually_ logged the position of each bulb so a computer could record the data (a servo-based system tracked the position of a pointer he placed on the screen and the computer logged the servo readings). A registration camera and projector ensure that adjacent frames of the film are in the same position - if you watch regular films (even Hollywood 70mm) where the frame edge is visible, you'll see the frames dance around a lot compared to a registration system. When you go through, say, 100 frames per second and then the entire several minutes of film, you've worn yourself out. However, the position info gave them accurate representations of the accelerations and therefore forces acting on the rig as it swung from lying on it's side to its final vertical position (not all rigs are built and launched this way - especially the big ones). He didn't use stereo lenses or film, but rather two cameras at right angles. With modern digital cameras, this could be done with a computer, especially if you rig different coloured LEDs and digitally filter out each one (say red, green and yellow and make the paddler wear blue in a blue kayak with a distinctly blue sea and sky). Digital cameras automatically self-register (fixed CCD position not moving film), but then the accuracy of the position info is dictated by the angular area of each pixel and the camera position relative to the paddler with appropriate corrections applied for perspective. If you have a 1000 pixel-wide field, and a field of view that's, say 3m wide at the paddler (ensure a 2.2m paddle stays in the field) then each pixel is nominally 3mm wide. You could probably interpolate the position to less that one pixel by watching pixel boundary crossings, but you'd also have to factor in what those Bayer algorithms do to physical vs logical pixels etc. Then you have to consider fixed camera position (fixed paddler; moving water) vs moving camera and then tracking the camera motion to convert relative to absolute motion etc. AFAIK, the software to do this you'd have to write yourself. Compared to that, the prospects of directly measuring the position sounds appealing. How to do _this_ remains to be seen :-) However, video based data logging is perfectly feasible. Mike *************************************************************************** PaddleWise Paddling Mailing List - Any opinions or suggestions expressed here are solely those of the writer(s). You must assume the entire responsibility for reliance upon them. All postings copyright the author. Submissions: PaddleWise_at_PaddleWise.net Subscriptions: PaddleWise-request_at_PaddleWise.net Website: http://www.paddlewise.net/ ***************************************************************************Received on Sun Jun 15 2003 - 15:13:22 PDT
This archive was generated by hypermail 2.4.0 : Thu Aug 21 2025 - 16:31:08 PDT