一种基于形态学的边缘检测算法

AN EDGE DETECTION ALGORITHM BASED ON MORPHOLOGY

在对含有噪声图像进行边缘识别时,为了提高识别精度,提出一种基于形态学的边缘检测算法。准备两种不同尺度的形态学结构元素,并对图像进行形态学降噪处理;用不同类型的形态学结构元素对处理后的图像进行边缘检测,获得不同结构元素下的边缘图像;根据每张边缘图像的信息熵来确定权值,并将这些边缘图像按照比例进行合成。这样,即使在有噪声干扰的条件下也能获得较为理想的图像边缘。实验结果展示了该算法相对于其他边缘检测算法的优势,突出其在保持图像边缘清晰的同时还具有较强的噪声去除能力,有力地说明了该算法的有效性和实用性。

In order to improve the accuracy of edge recognition of noisy image,this paper proposes an edge detection algorithm based on morphology.Two different scales of morphological structure elements were prepared,and the image was subjected to morphological noise reduction processing.Then,we used different types of morphological structure elements to detect the edge to get the edge image under different structural elements.Finally,the weights were determined according to the edge of each image information entropy,and these synthesized edge images were carried out in proportion.In this way,an ideal image edge could be obtained even under the condition of noise interference.The experimental results show the advantages of this algorithm compared with other edge detection algorithms,highlighting its strong noise removal ability while keeping the edge of the image clear,which strongly demonstrates the effectiveness and practicability of the algorithm.