一种新型的图像边缘检测技术

A new technology of image edge detection

针对传统的数学形态学边缘检测算法,提出一种新型的基于数学形态学的膨胀,腐蚀,开,闭等变换,采用不同方向不同尺度结构元素对图像进行边缘检测,大尺度结构元素清除噪声,小尺度结构元素获取完整的边缘。通过计算获取边缘的信息熵,自动匹配权值系数,进而得到最终图像边缘的完整信息。此方法能够有效地抑制噪声,保持边缘信息的完整性,精确性。

In view of the traditional mathematical morphology edge detection algorithms, this paper puts forward a new kind of mathematical morphology based on expansion, corrosion, opening operation and closing operation, such as the multidirectional multi-scale structure element is adopted to detect image edge, large scale structure unit is to remove noise and the small scale structure unit is to get completive edge. By calculating information entropy of the edge, the weight coefficient is adaptively determined, then achieving the final complete information of the image edge. The experiment proved that this method can suppress noise effectively and keep the integrity and accuracy of the image edge.