Study On Surface Model Reconstruction Technology Of Broken Skull Based On Reverse Engineering

The aim of this study is to adopt the technology and method of Reverse Engineering to make the Skull-Repair-Technology digital and improve the quality and efficiency of the repair technology. During this process, we can construct the Repair-Sheet that fits the skull surface and meets some certain continuity firstly, and then mold it. This method provides stable technique foundation for the rapid repair of skull, on the other hand, it can shorten the time of operation, reduce the risk of operation, and lighten the pain of patients. In addition, because of the use of digital produce technology, the manufacture cost reduces a lot as well as strong flexibility and high efficiency.During the study of the work, firstly, the author gains the data of broken skull through CT machine, and obtains a series of curves including some broken curves after distilling the data of the figure edges. Next, when designs Repair-Sheet digitally, the author links and integrates these broken curves, which is named Bridge-Link. Then, the surface of Repair-Sheet can be constructed by the linked and integrated curves. Last, the Rapid-Prototyping-Machine receives the data of the final surface of Repair-Sheet, and the sheet is manufactured.When constructs the Bridge-Link curves, in order to make these curves connect better to the curves which is at their two ends, the author designs an algorithm, through which these curves can be linked with G~2 continuity. At the same time, considering the feature of skull surface, the author also constructs a certain Bridge-Link curve that can be linked whit G~1 continuity. In order to test the theory and algorithm about Bridge-Link, a Bridge-Link system is designed. After analyzing the system and the result of test, it proves that the Bridge-Link algorithm can realize the skull figure edges curves linked perfectly. However, during constructing G~2 continuity Bridge-Link curves, the system needs two parameters (α and λ ), and the value of the parameters affects the result a lot, so there will have to do much work to solve these problem in the future.