图像方向性的条纹分析方法  

Stripe method to detect the image orientation

作　　者：庞韶宁[1] 李介谷[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200030

出　　处：《红外与激光工程》1999年第4期25-29,共5页Infrared and Laser Engineering

摘　　要：在图像分析方法研究中，提出了用条纹概念来分析图像，并成功地把这一理论应用于图像的方向性检测。文章首先从直观的条纹图像三维对比中，归纳出图像条纹概念的两个基本属性，然后构造了区别于传统的方向滤波方法的三角分形基迭代函数系统的图像方向提取模型，最后定义了图像的方向统计表征量。实验证明了条纹分析方法的高效性和实时性要求的独特性能。In this paper, stripes——a new image conception was proposed, which can be used in image analysis. Applying the theory in image feature detection, the orientation from the samples was detected successfully. Firstly, by the contrast of two stripe images in ( x,y,I ) space, two basic properties were concluded. Then, based on the triangle fractal base IFS, an image orientation detection math model was constructed. In the end, two variables to describe the image orientation were defined. The experiment indicates that the method is efficient and practicable. \;

关 键 词：条纹 分形 圆周采样 图像方向性 模式识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]