摘要
利用改进的BP神经网络模型导出了一种新的彩色图像边缘检测算法。为了充分利用图像中的颜色信息,在RGB彩色空间中通过欧式距离度量像素之间的差异获得灰度图像;为了降低训练样本的数量,将灰度图像二值化作为导师信号;针对传统的边缘检测算法容易产生边缘断裂、不连续等缺点,文章将动量法与自适应学习速率结合起来对传统的BP神经网络进行了改进。利用该方法对二值图像进行了边缘检测,实验结果表明,该方法对二值图像的边缘检测较传统的检测方法具有更好的效果。
This paper propose a new edge detection scheme for color linage based on improvea BP Neural Network. In order to fully take advantage of the color information of the image, the grayscale image is obtained by Euclidean distance between the pixels in RGB color space. The grayscale image is binarized to be taken as supervised samples to reduce the number of training set. Considering the shortcomings that traditional edge detection algorithms tends to lead to the discontinuity of detected edges, the method of momentum and adaptive learning rate are combined to improve the traditional BP neural network in the paper. Taking advantage of the method, the edge in a binary image is detected. Experimental results demonstrate that the method is superior to the conventional methods in terms of detecting the edges for binary image.
出处
《煤炭技术》
CAS
北大核心
2011年第10期154-157,共4页
Coal Technology