Sorry, i’m a window guy, i honestly don’t have the knoledge to code cross platform interface.
this said, except the interface (winforms) and camera acquisition(directshowlib),
the rest can surely be used in mono
the interface is fairely separated from the core engine in the code.
The core engine use interface to comunicate with the hardware, so verry little quantities of code will be needed to make it work on another platform or another harware.
ps: event the 3D rendering use interface,( for now it use an opengl port for c# : openTK)
I’m coding it as a hoobyist, but i’m a assiduous monotask hobyist
you can see the screenshots, the code is not in a so early state
I’m not happy with the interface, and i will rework it but the application is working.
the program is articulated around process :
a process is a list of task.
there are task for scan, load file, save file, filter, average, smooth, merge, mesh build, etc
a typical scan process is composed by these task:
scan => filter noise => smooth => merge lasers => buid mesh => save.
these typical process already work even if some task can be perfected.
If you want, you can just make a process with 2 task " scan" and “save” and work the result in another software (i only save stl for now, but a text pointcloud is a matther of 20 line of code).
but i want to make a “all in one” program and reconstruct the mesh (mainly for my intelect pride )
here is what i will do to finalyse the code.
find better filtering/average/smoothing algo.
better mesh reconstruction.
And a way to save load process (list of task, you can see them in eht screenshots)
add a system of plugins, so poeple can add custom task to a process.
as the calibration is realy not easy, i will surely do a “point-click helped” calibration wizard or so.
I’ve seen it, but when i’ve started, nothing was released yet. and as i’m a impatient guy…
i’ve take a look at the ciclop firmware, if all are so complete and well done, it will be surely be a reference.
I’m impatient to see the application that goes with it, and how they manage to remove noise, reconstruct meshes and do the calibration ( my 3 bigest problem ) i hope they have solutions and don’t rely on meshlab to reconstruct the stl.
