基于模糊熵和BP神经网络的彩色图像边缘检测

Color image edge detection based on fuzzy entropy and BP neural network

在RGB颜色空间中,分别提取R、G、B三个分量并计算R、G、B三个分量的组合V,通过引入模糊熵,构造出4个基于模糊熵的信息测度分量来定量描述彩色图像的边缘特征,并将4个测度分量组成一个整体的特征向量,计算训练图像的特征向量作为样本对BP网络进行训练,然后将训练的BP网络直接用于边缘检测。该方法充分考虑了颜色空间中各颜色分量以及它们之间的相关性;BP网络的结构和训练都比较简单;实验表明,改进方法具有较强的细节保持能力,对弱边缘具有较强的检测能力。

An improved method of color image edge detection based on fuzzy entropy and BP neural network is presented. Firstly, the R, G、B of the color image are extracted and the combination of the three components is computed in the RGB color space.Through quoting fuzzy entropy, four information measures fuzzy entropy-based to describe the edge feature of color image are constructed and a feature vector is composed of the four measures components.Then through the training with the feature vector samples calculated from training images,the BP neural network acquires the function of a desired edge detector.Finally, the trained BP neural network is used for edge detection directly.The method is fully considered between the color components and their correlation in the color space.Besides, both the architecture and the training of BP neural network are simple.The experiment＇s result proves that this method has the strong retention capacity of details and is more sensitive to weak edge detection.