Nowadays, the large-scale development of IT, machine vision is becoming an increasingly important element of our life. Computer vision systems are available in several part of IT. Thus they are using in radiotherapy too.
In my thesis I used AlignRT camera system datas which manufactured by VisionRT. AlignRT is an intelligent camera system. It is using two or more 3D stereo camera to reconstruct pointclouds. My thesis main goal is finding the best algorithm for point cloud registration.
The introduction part of the thesis will introduce readers to radiotherapy treatments and the widely used AlignRT system. This part will define the main problems and give a brief description of used softwares.
Then I will show the local algorithms, Iterative Closest Point and it variants. With the measurement results I want to find the best local algorithm and their problems.
The solution for local algorithm problems, can be global registration. I will present some global registration algorithm with measurements. Then I will close with algorithm conclusion.
I will present two possible processing flow solution for the patient setup and the patient positioning problems.
And the last part of my thesis will show the console application parts. This application uses the most algorithm which I mentioned before.