摘要
针对图像进行准确数字化描述是计算机图像处理的关键问题。在图像特征提取过程中,采用链码表和线段表描述特征向量的数据结构。首先进行轮廓跟踪,用行扫描得到图像轮廓起点,然后采用链码跟踪技术得到封闭的轮廓信息---链码表,直到所有的轮廓跟踪完毕为止。再通过线性转换得到线段表,最后根据链码表和线段表分别求出周长和面积等特征。此方法已在Visual C++平台实现,经验证,采用链码表和线段表两种结构可准确求出图像几何形状特征。
In view of the key point of exact digitalized description of image in image processing, chain codes table and line segment table are adopted to describe data structure of feature vector, In the process of image feature extraction, adopting contour tracking firstly, contour starting point is obtained with line scanning, then closed contour information---chain codes table is obtained with chain codes tracking technology, until all contours are tracked. Line segment table is obtained by linear cooversion. Finally, features such as perimeter, area and so on are computed based on chain codes table and line segment table. The algorithm has been realized with Visual C++ and the results show that adopting chain codes table and line segment table can compute image geometry shape feature exactly.
出处
《辽宁工程技术大学学报(自然科学版)》
EI
CAS
北大核心
2007年第2期257-259,共3页
Journal of Liaoning Technical University (Natural Science)
基金
辽宁省教育厅A类基金资助项目(20243303)
沈阳市科技局基金资助项目(20020256)
关键词
链码表
线段表
特征提取
chain codes table
line segment table
feature extraction