【作者】 蔡占川； 孙伟； 毕宁； 齐东旭；

【Author】 Cai Zhan-Chuan Sun Wei BiNing QI Dong-Xu School of Information Science and Technology, Zhongshan University, Guangzhou, 510275 Faculty of Information Technology, Macao University of Science and Technology, Macao School of Mathematics Computational Science, Zhongshan University, Guangzhou, 510275

【机构】 中山大学信息科学与技术学院； 中山大学数学与计算科学学院；

【摘要】 为了探索有效识别物体形状的新方法,引进一类正交完备的分段k次多项式系统(简称U-系统);U-系统是一类属于 L2[0,1]的正交完备分段k次多项式系统;该系统下的U级数展开式具有良好的平方逼近及一致逼近性质。基于U-系统理论, 提出了U描述子的概念,并且给出了U描述子的性质并在理论上予以证明。为了更好的识别物体形状,对U描述子进行了归一化,同时在理论上证明了归一化U描述子具有旋转、平移、尺度大小等不变的性质。实验表明,归一化的U描述子能够高效、准确的识别物体的形状,与Fourier描述子相比,具有更好的识别率。更多还原

【Abstract】 In order to probe into the recognition of object’s shape in effect, a class of orthogonal complete piecewise k-degree polynomials (so-called U-system)is introduced. U-system is a class of orthogonal complete piecewise k -degree polynomials in L2[0,l] .The expansion in U-series has advantageous properties for approximations in both quadratic norm and uniform. Based on U-system, A new U descriptors is defined, and present the property of U descriptors and its proof. To recognize better object’s shape, A new normalized U descriptor is defined, which is invariant in the rotation, translation and scale transform. The experiments show the accuracy and efficiency of new normalized U descriptors in shape recognition and better recognition ratio.更多还原