论文部分内容阅读
基于 Marr边缘检测算子 ,以高斯函数的一阶和 2阶导数作为小波 ,提出具有 2个可变参数的多尺度图象边缘检测方法。小波分解的级数控制观察距离“调焦”,改变 Marr算子中的标准差σ、可实现在同一分辨率级上所要求检测边缘的细节程度。本文所提出的多尺度边缘检测方法适应于计算机视觉分级识别中边缘检测的要求。
Based on the Marr edge detection operator, the first and second order derivatives of Gaussian function are used as wavelet, a multi-scale image edge detection method with two variable parameters is proposed. Wavelet decomposition of the series control to observe the distance “focus” and change the standard deviation σ in the Marr operator to achieve the level of detail required to detect the edge at the same resolution level. The proposed multi-scale edge detection method adapts to the requirements of edge detection in computer vision classification recognition.