| Based on the Computer Control System of Setting Fracture Manipulator, utilizing the theories of image processing, feature extraction and expert systems, the Expert System for Intelligent Fracture-Resetting Robot is a new kind of intelligent equipment which can include experience of many physic experts and help the doctors to finish auto-resetting fractures. The system can not only avoid doctors from the damage of X-rays by isolated operation, give a permanent and stable resetting force and register the real-time resetting data, but also recognize the fracture feature by image processing and finish the fracture resetting automatically based on the knowledge base of the expert system. Consequently, the quality of the fracture treat can be enhanced and the range of no-cut resetting can be enlarged.This thesis elucidates the whole design ideas of the Intelligent Fracture-Resetting Robot and presents the working of the control subsystem and the remote subsystem. We present and exploit an embedded expert system unit aimed at the problem that the expert can not handle CLIPS kernel directly. The unit can realize an all-purpose fuzzy inference mechanism and an interpretation mechanism by preprogramming in CLIPS kernel. The knowledge base is founded by the expert. We build a knowledge acquisition which has a unique dictionary. The unit also has a mechanism through which it can be called by any other applications directly, and can realize different functions for different objects by loading different knowledge bases.To achieve the purpose of auto-resetting fractures, this thesis discusses the scheme of processing and extraction about the fracture X-ray images. The original gray image is disposed by a serial of processing like Sobel's edge detection, image bi-valuing, median filtering and edge thinning. Finally it can be translated into an edge binary image whose width is only one pixel. Based on the features of the femoral neck fracture and correlative physic knowledge, we present fracture features of two angles and two distances. To make an extraction target of these features, the fracture edge image's contour should be tracked, and the original points and lines will be extracted by some methods like the improved Hough transform. Then we can get the final fracture features by geometry calculation.In the end, considering the practical application of the Expert System for Intelligent Fracture-Resetting Robot, we discuss the instauration of the knowledge base for femoral neck fracture and introduce the exploitation of the simulate software for the expert system. We also analyze the constituting of the remote system, the mode of the data transferring and the result of the internet testing. |
