Surface of Revolution Analysis bugs

[Getarq]

Hi Daniel, first of all congratulation for this amazing software, this is really useful, so thanks a lot for this.
We are trying to use the software to make some oil Tanks analysis.
We are working with the Surface of Revolution Analysis plugin, this plugin is really good, but we found some bugs that are a bit annoying.
1- we work with the unroll projection to show the critical points obtained after run the surface analysis, the start point for the unroll is the opposite start point of the 2d deviation map, this make this image useless because change the interpretation of the images respect angles and height. For now I'm remaking this image with my own grid in a cad software using the unroll projection image.
2- The ASCii grid exported from the 2d distance map, has the height parameter inverted. The zero value is in the top of the grid, And the top value of the tank in this case is in the zero value of the grid. So when we try to produce some graphics of Verticality using the top and zero values to compare them with the vertical profiles, nothing matches.
3- No, because we are using the unroll projection for showing the critical deviation points, and this unroll from another start point, all the angles and information obtained from the surface revolution analysis doesn't match this image, so We must make complex modifications to the data obtained int the vertical and horizontal profiles and in the ASCii grid.

Hope you can help, or give me any advice.

Again, thanks a lot for all your amazing work
Greetings
Diego

[daniel]

1. In fact the SRA plugin has a 'CCW' checkbox that allows you to choose in which side the shape should be 'unrolled'. Or did I misunderstood something?

2. It may depend on the order in which the profile is defined (or the position of the origin). Have you tried to play with these 'parameters'?

3. The qSRA plugin has been developed by a private company (for their own usage). They didn't use the same angular 'zero' position as the 'Unroll' tool. However, I believe a simple rotation around the vertical axis (of a multiple of 90 degrees) should do the trick, isn't it?

Last but not least, if you want me to 'debug' anything, I'll need some data to reproduce these issues. You can contact me via email (cloudcompare [at] danielgm.net).

[DingoDom]

Hi Diego,

I'm trying to recreate what you have done for a university project. I have a cloud scan of a tank and am trying to compare it to an ideal drawing. Could you please give me some direction on how you came up with your images. Particularly the one with the measured deviation.

Thanks
Dom
Brisbane Australia

[daniel]

Everything's here: http://www.cloudcompare.org/doc/wiki/in ... s_(plugin)

And if your shape is a cylinder, you can now create a cylinder entity (with the Primitive Factory) and use it as reference (it's simpler than defining the profile manually).