Abstract :
The slicing contour is one of the most common used original information in reverse engineering. It is a dot matrix image of prototype’s cross section obtained from prototype being measured in certain direction, the slicing layers are parallel to each other and the spacing between them is teeny. The technology recognizing modeling feature based on series of slicing contours information is studied in this paper. The modeling feature defined in this paper means the solid and surface which can be constructed one time using single modeling method in universal modeling software environment. For any features, no matter what the direction of slicing layer is, the neighboring slicing layer shape is invariable, or it changes gradually, it is impossible for it to change suddenly (slicing layer producing sudden change is certainly the joint of two features). To recognize feature according to the dot matrix image of slicing contour, it is important to know the extent of likeness between the neighboring contour dot images. It isn’t necessary to know exactly what the contour shape is. Therefore when matching neighboring contour dot images, it needn’t to compare dot by dot, it only needs to compare whether the shape, position and size of the neighboring contours are alike. In this paper, the mathematics expression of slicing contour dot image is proposed in order to describe the shape, position and size of the contour, the feature recognition theorem is given and the feasibility of the method is shown by using an example. This method is simple, practical and efficient.
