基于改进的形态学灰度图像边缘检测算法

Improved Edge Detection Algorithm of Grayscale Image Based on Morphology

摘要:

针对传统形态学在边缘检测算法上存在的检测边缘错位、较粗等缺点, 在传统形态学的基础上, 提出了一种具有方向估计性的梯度算子。该算法通过改进的梯度算子及多结构元素模式来获取梯度方向, 并沿边缘梯度方向进行非极大值抑制, 以此细化边缘宽度, 最终获取图像边缘。实验结果表明, 相对传统形态学边缘检测算法, 该算法具有更精细的检测结果与较强的抗噪能力。

Abstract:

In order to solve such problems as edge dislocation and coarser in traditional morphological edge detection, a gradient operator with the property of direction estimation based on the traditional morphological is proposed. In this algorithm, gradient direction is obtained by using the improved gradient operator and the mode of multi-structural elements. To refine the edge width and ultimately get the image edges, non-maximum suppression along the edge gradient direction is adopted. Synthesis results confirm that compared with the traditional morphological edge detection algorithm, the proposed algorithm can get better test results and stronger anti-noise capability.