基于形状特征的二维旋转尺度不变相似度量

目标物体的旋转及尺度变化在目标识别过程中是普遍存在的 .这里给出了一种新的特征提取方法及其相似度标准 ,通过提取物体形状中具有二维旋转尺度不变性的信息 ,来实现对二维变化目标的识别

Rotation and scale changing usual take place during pattern recognizing. This thesis propose a new method of abstracting character of object and its resemble criterion by abstracting characters containing information independent to 2D rotation and scale of object, apply to recognize changing object in 2D.