CAD/CAI中的图形识别研究

Figure identify in CAD/CAI

图形的自动识别是计算机辅助设计和辅助教学研究中的一个难点,识别图形的重要判据是两图的特征重合率.本文对图形的基本要素之一———直线的相似判别进行了研究,霍夫变换方法常应用于图形判别和图像识别,霍夫变换是将图形空间的点根据某些约束条件拟合成线段,点可以用通过该点的线段集合来描述.本文最后利用了聚类方法中的最小距离法进行了两个图形特征相似性的判别,通过最小距离法可以进行图形正确与否的判别.

Figure identify is one of most difficulty problems in CAD/CAI research. The importance criterion of figure identify is the ratio of feature superposition for two images. Identify of lines similarity which is one of essential element in figure identify is researched in the paper. The method of Hough transform is often applied to figure identify and image recognition. Hough transform describes the point in graphics space is bound by synthetic line under certain conditions. The point is described by the aggregation of lines. Clustering method is used here in the end. The minimum distance law is used in a similar distinguishing between two characteristics of images. The figure is identified right or not by using method of minimum distance.