摘要
本文描述了一种角点的形状分析方法。它选用相邻角点与形心所成直线之夹角,以及角点到形心之间的归一化距离作为特征,并证明了它们的平移、旋转和比例变化不变性。为了提高特征匹配的精度,本文采用动态分析的方法,将其应用于飞机目标识别之中。实验证明,这二组特征有较好的分类能力。
This paper describes a shape analysis method based on corners. Features used in the method are the angle of two lines from centroid to neighbouring corner points and distances between centroid and the corner points. It is proved that these features are invariant to translation, rotation and scale transformations. Dynamic feature matching is used in this paper to improve the recognition and classification precision. The method is used in the recognition and classification of aeroplanes. Experiments prove that these two classes of features have good results for aeroplane recognition and classification.
出处
《数据采集与处理》
CSCD
1992年第3期182-186,共5页
Journal of Data Acquisition and Processing
