中文摘要：

Radon变换已经成为一种图像处理的基本方法,成为图像重建技术的基础,可以将直线与变换域的点相对应,将数据降维,便于进一步处理。但经典的Radon变换方法对于非直线情况处理效果不好,图像中的多样性信息无法充分利用。文中基于广义Radon的思想,考虑到识别目标的多样性,提出了一种利用物体形状信息的比较通用的目标识别方法,图像点映射改善了实际运算方面的性能,当目标信息比较准确时,新方法的效果比较理想。根据该方法进行的相关实验,也得到了较好的印证。

英文摘要：

The Radon transformation,as one popular essential imagery processing method,provides the theoretic foundation for image reconstruction technology,mapping a straight line to a point of transformed domain and get sparse representation,good for further processing,but the transitional method does not work regarding the non-straight line processing,neglecting image multiple information. Based on the generalized Radon transformation method,in view of multiple information of object＇s shape,a new general target identification algorithm is proposed. The image point mapping （IMP） improved the actual practical performance, and it makes a good effect when accurate information of object is available. Corresponding experiment carried on which according to this method, also got a good verification.