摘要
提出了一种适合于中药材贝母粉末显微图像的纹理分析方法,即综合运用分形维数、二级小波变换和灰度—梯度共生矩阵3种方法分析纹理图像,共提取了8个特征量用于模式识别实验,得到了较好的识别效果,表明模式识别技术是可行的。
An effective texture analysis way that adapts to Chinese traditional medicinal material Beimu powder micro-image is presented. This way combines fractal dimension, two-level wavelet transform and gray-gradient co-occurrence. Eight character parameters are singled out and applied to pattern recognition experiment. The better classification is obtained. The experiment results show that this pattern recognition technology is viable.
出处
《计算机工程与设计》
CSCD
2004年第10期1676-1677,1684,共3页
Computer Engineering and Design
基金
辽宁省自然科学基金(972147)
辽宁省教育厅基金(20182241)
