摘要
提出了一种应用混合神经网络进行颗粒图像边缘检测的方法·边缘候选图像是通过采用基于灰度极小值算法提取边缘候选像素获得，神经网络以边缘候选图像中的边缘候选像素及其邻域像素的二值模式作为训练样本·对经过噪声污染的图像进行实验表明，该方法获得的边缘图像封闭性较好、边缘描述真实，适用于颗粒图像的边缘检测·
An edge detection method is proposed to examine globular materials using hybrid neural network. A logical judgement algorithm was used to get edge candidate images. Edge pixel candidates and their neighbor pixels compose the binary samples of the hybrid neural network. Tests on the images corrupted with Gaussian noise show that the images segmented by this method have good edge closedness and true edge. The images are suitable for the segmentation of cumulate particle image.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
1999年第2期126-129,共4页
Journal of Northeastern University(Natural Science)
基金
辽宁省科学技术基金
辽宁省教委课题
