基于信息测度和支持向量机的图像边缘检测

Application of information measure and support vector machine in image edge detection

将信息测度和支持向量机结合在一起,提出了一种新的图像边缘检测方法(information measure and support vec-tor machine edge detection method,ISEDM).首先,基于数学测度概念构造一个描述边缘点信息测度的特征矢量,该矢量由邻域一致性测度、方向性信息测度和梯度分布3个特征分量组成,然后运用支持向量机对特征矢量数据集进行训练和分类,实现了对边缘点的检测.实验结果表明,对于含有加性噪声、乘性噪声等图像的边缘检测,ISEDM能够有效地抑制噪声,较多地保留图像边缘的细节信息,边缘图像锐利而清晰.

A novel method for image edge detection is presented based on information measurement and Support Vector Machine, which is called ISEDM （Information measure and support vector machine edge detection method）. At first, a vector is constructed to fully describe a edge point information measure, which includes neighborhood homogeneity information measure, orientation information measure, and gradient strengths. Secondly, SVM is applied to train and classify the set of feature vectors, so that the edge of the image is detected. The experimental results show that ISEDM can not only effectively reduce the noises of the image, but also can precisely detect the edge-position, and keep the image edges＇ details well.