摘要
图像特征点检测是图像匹配、目标识别以及运动估计等领域的一项关键技术。本文对图像轮廓二维信息进行降维处理,提出了一种特征点质量评价因子。利用该因子并结合文中给出的特征点提取准则对图像轮廓链码进行分析,提取特征点。该方法避免了常规的基于链码的特征点检测方法中曲率的计算,提高了检测速度。试验证明该方法具有较好的实时性和定位精度。
Image feature point detection is a key technique in the domain of image matching, targets identification and motion estimation. In this paper, a quality factor is presented by changing 21） image contour information into 1D. This quality factor and some criterions about feature point extraction mentioned in this paper will be used simultaneously to analyse the chain code of image contour, thus complete feature point extraction. This method avoids the computation of curvature in conventional chain code based method; thus, detective velocity is improved. Simulation results show this method has better solution in terms of real-time and location precision.
出处
《微计算机信息》
2009年第25期123-125,共3页
Control & Automation
关键词
质量因子
链码
轮廓分析
特征点检测
Quality factor
Chain code
Contour analysis
feature points detection
