# Share a workflow for RTK beginners

**URL:** <https://discourse.itk.org/t/share-a-workflow-for-rtk-beginners/7439>\
**Category:** Community\
**Created:** [February 19, 2025, 6:33am UTC](https://discourse.itk.org/t/share-a-workflow-for-rtk-beginners/7439 "2025-02-19T06:33:02Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![zhangxiang](https://discourse.itk.org/letter_avatar_proxy/v4/letter/z/c77e96/32.png) [@zhangxiang](https://discourse.itk.org/u/zhangxiang)\
**Post date:** [February 19, 2025, 6:33am UTC](https://discourse.itk.org/t/share-a-workflow-for-rtk-beginners/7439/1 "2025-02-19T06:33:02Z")

</div>

This workflow can be divided into three steps: calibration, transfomation of geometric parameters, using RTK

1.calibration  
method1 : DLT and optional optimization on reprojection errors. If you want to understand DLT, detailed process can be seen in paper[1].  
method2: use a one-dimensional phantom. see paper[2,3,4]. wish you are good at mathematics.

1. transformation of geometric parameters  
RTK model[5] is different from a pinhole model( DLT). After obtaining geometric parameters of a pinhole model (the Z axis of the pinhole model here is the rotation axis of CBCT system), you should transform what you get into RTK model.  
I cannot share the code for some reasons. Learning Chaper 1,2 in Robotics may help a beginner to understand transformation matrix. A link[6] for Chinese.

2. RTK  
CBCT system I use is a object-rotated CBCT. 2D X-ray image is 768×768 with 0.56mm/pixel. The rotation direction of turntable is clockwise around the Z axis. Gantry angle is not necessary.

.\rtksimulatedgeometry.exe -n -o ‘’‘////xml’ --sdd --sid --source\_x --source\_y --proj\_iso\_x --proj\_iso\_y --out\_angle --in\_angle

.\rtkfdk.exe -g -o -p -r “.\*.dcm” -v “on” --neworigin “-214.76,214.76,0” --newdirection ‘‘1,0,0,0,-1,0,0,0,1’’ --newspacing “0.56,0.56,1” --hardware “cuda” --spacing --dimension

newdirection is ‘‘1,0,0,0,-1,0,0,0,1’’ because the v diretion of RTK model is opppsite to the v direction of a 2D image.

214.76=(768-1)/2\*0.56

```
 [1] New Calibrator with Points Distributed Conical Helically for Online Calibration of C-Arm
[2] Analytic method based on identification of ellipse parameters for

```

scanner calibration in cone-beam tomography.  
[3] A graphical method for determining the in-  
plane rotation angle in geometric calibration of circular cone-beam ct  
systems  
[4] An analytical geometric calibration method for  
circular cone-beam geometry,  
[5][RTK: RTK 3D circular projection geometry](http://www.openrtk.org/Doxygen/DocGeo3D.html#RM)  
[6][台大机器人学之运动学——林沛群（含课件+书籍）\_哔哩哔哩\_bilibili](https://www.bilibili.com/video/BV1v4411H7ez)