摘要
对粉煤图像和块煤图像灰度相关矩阵各统计量的数值进行了极差正规化处理,并分析了其统计量的分布特征.使用概率神经网络对粉煤图像和块煤图像进行了识别仿真.实验结果表明,用灰度相关矩阵各统计量作为粉煤图像和块煤图像的识别特征,成功地识别出了粉煤和块煤的图像.
The normalization values of texture statistics of gray level correlative matrix were given, which were taken from the smashed-coal-images and block-coal-images. The distribution feather of statistical variables was analyzed. Recognizing the smashed-coal-images and block-coal-images was simulated with a probabilistic neural network. The experiment results show that the statistical variables of the gray level correlative matrix act as the recognizable feather, and the algorithm can recognize the smashed-coal-image and block-coal-image successfully.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2007年第11期1206-1210,共5页
Journal of China Coal Society
基金
教育部博士点基金资助项目(20050290010)
北京市教育委员会共建经费研究生教育资助项目
关键词
概率神经网络
粉煤图像
块煤图像
灰度相关矩阵
probabilistic neural network
smashed-coal-image
block-coal-image
grav level correlative matrix
