摘要
该文描述了双高斯差(DifferenceofGaussian,以下简称DOG)函数和高斯函数及高斯-拉普拉斯二阶微分函数之间的关系原理,提出了建立在DOG模型基础上的新角点检测算法。实验结果表明DOG角点检测器具有检测范围宽、抗噪声能力强、检测方式灵活等优点,具有良好的角点检测性能。
The principle of relationship between difference of Gaussian(DOG)model and Gaussian kernels is described in this paper.Based on the difference of Gaussian(DOG)model,a new corner detection algorithm is presented.Experi-mental results show that DOG model-based detector has outstanding characteristics such as wide detection spectrum,high performance to anti-noise and flexible detection form,et al.The detector has the efficiency for corner detection.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第11期87-90,99,共5页
Computer Engineering and Applications
