It was shown that the matching of the CT scan models with acquired models can be successfully carried out using the proposed procedure and its associated mathematical analyses. Problems encountered with camera calibration, robot arm calibration, correspondence of stereo images were shown to present major difficulties, but were proven to be surmountable using proposed solutions. These algorithms were demonstrated using an experimental computer code written in MATLAB® and were validated through experimental tests.
Although the camera calibration was not successfully carried out in the experimentation, and thus the procedure was not completely simulated, it is believed that with better quality cameras, the 3D acquisition can be effectively and accurately performed. This proposition is supported by the different tests performed on synthetic images. As for the matching and correspondence problems, novel solutions, which rely on 3D-model fusing and color discrimination respectively, were proposed and successfully tested.
In conclusion, robotic assisted orthopedic intervention should result in and increased 3D visualization, and thus better preoperative planning, and in an increased accuracy in the delicate, yet firm drilling and cutting actions. The additional requirements for robotic assisted orthopedic interventions over manual ones are a 3D-acquisition system that offers high precision, a robotic manipulator, and an appropriate CAD system.