论文部分内容阅读
提出一类渐近收敛于高斯函数的导数的中心B样条二进小波,用以实现多尺度Canny最优化准则.在应用Canny准则的实际计算中,必须对高斯函数的导数进行截尾.文中通过分析开窗截尾高斯函数的导数,对滤波器性能和计算的快速和有效性的影响,提出了利用具有紧支集的一类中心B样条二进小波作为滤波器,一方面可以近似满足Canny提出的最优滤波器的要求;另一方面,可以进行快速有效的多尺度边缘提取
A class of central B-spline binary wavelet converging asymptotically to the derivative of Gaussian function is proposed to realize the multi-scale Canny optimization criterion. In practical calculations using Canny’s criterion, the derivative of the Gaussian function must be truncated. In this paper, by analyzing the derivative of windowed truncated Gaussian function, the filter performance and the calculation of the rapid and effective, the use of a class with tight branch B-spline bilevel wavelet as a filter, on the one hand can Which can approximately meet the requirements of Canny’s optimal filter. On the other hand, it can perform fast and efficient multiscale edge extraction