PTX transformation question

Development Asset Creation
,
1

Hello,

I’m trying to use custom created PTX files in RealityCapture. Problem is there is quite little information about PTX format and RealityCapture doesn’t seem to show scanned point clouds before processing. 

Could someone explain following questions about the PTX header?

st1 st2 st3 ; scanner registered position 
sx1 sx2 sx3 ; scanner registered axis ‘X’ 
sy1 sy2 sy3 ; scanner registered axis ‘Y’ 
sz1 sz2 sz3 ; scanner registered axis ‘Z’ 
r11 r12 r13 0 ; transformation matrix 
r21 r22 r23 0 ; this is a simple rotation and translation 4x4 matrix 
r31 r32 r33 0 ; just apply to each point to get the transformed coordinate 
tr1 tr2 tr3 1 ; use double-precision variables 

in most usages I have found that st and tr vectors are the same. And sx, sy, sz are same as r matrix.

What meaning has this duplicity?

Is matrix R rotation from camera space to world space?

Should the points be already transformed by this transformation?

What types of transformations are done to the points during import and processing with RC?

2

Hello David,

RC ignores the first 4 lines of the header during import. It uses rotation and translation matrices to transform individual points, that are in scanner coordinate system into global coordinate system.

Rotation and translation matrices define the transformation from scanner coordinate system into global coordinate system.

No, individual points need to be in the scanner coordinate system.

RC uses only rotation and transformation matrices defined in ptx file.