基于多尺度形态学的边缘检测算法的改进

摘要

一幅图像的大量信息都是由图像边缘提供的，在图像处理中，对于复杂图像很难通过单一的结构元素用数学形态学的膨胀、腐蚀、开、闭等变换以及它们的组合得到较满意的检测图像；利用多元素，对形态结构大小尺度进行调整，从而在噪声存在的条件下得到较理想的图像边缘；实验结果表明，与传统的边缘检测算子比较，方法计算量小，抗噪性好，可以适应不同类型的图像边缘检测需求。%Ab A lot of information of a digital image is provided by image edge, in the processing of an image, as for a complex image, a satisfactory detection result is difficult to be obtained by using the transform such as dilation, erosion,opening, closing and so on and their combination of mathematical morphology through a single structure element. This paper uses multi-element to adjust dimension of morphological structure so that relatively ideal image edge can be gained under the existence of noise. Experiment results show that, compared with traditional edge detection operator, this method needs small computation, has better noise immunity and is suitable for the demand of different image edge detections.