图像边缘提取中的一种动态阈值获取法

图像边缘是目标识别，图像理解及图像检索中常用的一个重要的特征。随着多媒体技术的发展，图像库变得越来越大，使得特征的自动提取显得越来越重要。然而在大图像库中，每幅图像的目标和背景间的变化均不相同，自动提取边缘过程中，不同图像边缘的阈值选取非常关键。为实现大图像库中像边缘的自动提取，阈值的动态调整是很重要的，本文提出的动态阈值调整法，能够自动提取各各变化不同的图像边缘。

A METHOD TO GET THE DYNAMIC THRESHOLD OF IMAGE EDGES

Image edge is an important feature in object recognize, image understand and image retrieval. As the development of multimedia, the image database is becoming larger and larger, and it is becoming more important to get the features automatically. But in large image database, as the changes between object and background are various in every image, it is difficult to get the threshold dynamically. In this paper, A method to get the dynamic threshold of all image edges is presented. It can automatically extract edges of every image in the image database.