Sunday, January 9, 2011

Chosen Building: Mesh Reconstruction.

Process of Reconstruction:

Stage1: Import into Meshlab.
The larger scale of the chosen building consisted of more details which resulted in a more segregated distribution of feature points. This is more noticeably the case around areas which are clad in glass that has a high reflectance.


 Stage2: Cleaning unnecessary vertices.
 As I intend to only incorporate the single column into the creation of my artwork, I have removed the building facade on adjacent sides, this is particularly easy when I selected vertices based on texture colour. However it will be more convenient if there were some sort of dropper tool which can help pick out the texture colour as oppose to manually trial and error.


Stage 3a: Attempted reconstruction 1.
Even though I had anticipated a straight forward outcome after gaining most experience with previous reconstruction of the bag, however the first result was not a sucessful one as I had intended for the column to be a sharp and linear element. The reconstruction had become one giant ball of  blob due to the scattered vertices that were not properly cleaned up and created a polygon connection of the two ends.


Stage 3b: Attempted reconstruction 2.
After another attempt at cleaning up any vertices that were accidentally left out previously, the reconstructed mesh has began to take shape of the column, however blob still exists on the top of the structure where the timber members connect to the glass roof due to the lack of feature points from the aerial view.


Stage 3c: Reconstruction.
For the final attempt at create an accurate reconstruction, I removed most of the glass roof to give more emphasis to the branching out of the timber elements on the top of the column.

This failed at first as it ended up becoming a non-watertight mesh, by removing the additional faces that have no thickness, I will try to enclose this mesh in 3ds max with the boolean operation.


Stage4: Import into 3ds max.
As more clearly shown in a render, the lack of feature points made it difficult to recreate the whole column, especially the back which were heavily shadowed.


Stage5: Comparison.
In contrasting with my chosen object, it is pretty clear that the photosynth result of the scientia column does not achieve a clear level of representation, and will be even harder to identify when 3d printed without a texture, nevertheless, having been able to retain the general shape of the column as a sharp object allowed me to pierce through the bag, and its irregularity in form also create a merge between what is natural and artificial.



No comments:

Post a Comment