摘要
岩石节理裂隙形状复杂,图像中含有很多噪声。而使用统计模式识别方法在分割图像时,可以首先使分类器学习图像中不同样本的特征,进而利用这些特征对图像中每个像素进行分类,实现分割。在设计统计模式识别的分类器时,我们提出使用基于核函数Fisher判别法构造分类器。使用该方法可以将图像高维的属性空间上的非线形判别转化为图像特征空间上的线形判别,而不需要知道从属性空间到特征空间的具体映射形式。通过对岩石节理裂隙图像分割实验表明,该方法能达到较其他方法更好的分割效果。
Rock joint network is complex.There are some noises on a rock joint image.But the classifier can study features of different samples in an image when we use statistical pattern recognition method to segment the image,Then each pixel in the image is classified based on these features.As a result,rock joints are segmented from the image. When we design the statistical pattern recognition classifier,Kernel Fisher discrimination is used.Kernel Fisher discrimination can transform nonlinear discrimination in an attribute space with high dimension into linear discrimination in a feature space with low dimension.And we need not know the detailed mapping form from attribute space to feature space in the process of transformation.We conduct some experiments on rock joint images.The results show that this method gives better segmented images for all testing images.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第10期213-215,共3页
Computer Engineering and Applications
基金
TRUE(TheTracerRetentionUnderstandingExperiments)
由瑞典SKB
欧盟联合支持
关键词
核函数
FISHER
判别法
统计模式识别
图像分割
岩石
节理裂隙
Kernel function, Fisher discrimination, statistical pattern recognition, image segmentation, rock ,joint