Analysing a photo of a mechanical cam

[peterAustralia]

hi all

I have an image of a mechnaical cam from a linotype printing press.

I want to get angular dispalcemetns, angular velocities and angular accelerations of this cam from the image.

Is this possible. Does anyone know any software where I can get an image and then get a set of coordinaates of the perimeter for analysis. My rough guess says that if I had 500 or so coordinates I could get a very good approximation of the cam.

so is there any software out there that can get me the coordinates I seek. Is there other software in existence which I can use to analyse the cam. Since this is a 2D image, perhaps a CAD program could analyse the coordinates if they were in dxf format of possibly csv?

my background is coding and web design. This is for a friend who has been working on this 'project; for almost a year and going mad. (yes he has a lot of spare time!). Thus the finer points of mechanical cam analysis and software are actually lost on me

thanks in advance if you are able to help
n peter evans

 

[Task]

...perhaps a CAD program could analyse the coordinates if they were in dxf format of possibly csv?

I'm unsure about the analysis part, but http://www.alibre.com/xpress/software/alibre-design-xpress.asp" will import 2D .dxf 's into 3D & its free.

GL hope it helps.

 

[FredGarvin]

I don't see how you are going to get all of that information from one image of the cam. Do you have multiple images that are time coded or something along those lines?

Is this cam in operation right now? If it is, is there any way one could install even a simple prox probe to take some measurements?

I have, from time to time, used a simple scale with an image to reverse engineer something, but you have to make sure that the scale the photo was taken at is reliable and verifiable. If you had that, then pretty much any 2D CAD package will import an image to work on top of.

 

[brewnog]

Tough one. I'd probably be getting the image scaled up, getting it on a drawing board, and using calipers to take measurements at indexed points from the centre at as many positions as possible. Then plot out distance against angle in Excel or some such.

The rest is just a case of some mathematics, but I'm not sure how accurate your initial measurements would be.

 

[jaap de vries]

Check for "spotlight", freeware created by NASA. Do you need to to real time analysis??