论文部分内容阅读
目的:编程实现对防风药材中多种组织结构的量化表征,为定量描述中药材显微鉴定特征提供技术支持。方法:根据药材显微图像中主要形态结构的特征,利用数字图像处理技术和采用MATLAB软件编程的方法,进行防风药材横切面各视野显微图像的摄取和编码,采用模板匹配法对各视野灰度图像进行配准并拼接,用Canny算子对拼接图像进行边缘检测,通过形态学基本操作对拼接图像进行图像二值化以及分割各种显微特征目标物和分析其大小、形态、数量,并采用系统聚类法对油管和导管的分布进行表征;将提取的四个产地防风药材的数据进行判别分析。结果:从样品图像中分割出了防风药材的木栓层、韧皮部、木质部、油管、导管等横切面中的主要显微特征;从分割的图像中各提取了20个参数,经逐步判别分析筛选出12个参数,以此对4个产地防风药材样品在组织形态方面的差异进行判别分析,总正确率为91.4%。结论:该方法能较准确地描述4个产地防风药材横切面组织结构的细微差别,这对不同产地防风药材的显微鉴别具有一定的参考意义。
OBJECTIVE: To program and quantitatively characterize various tissue structures in windproof medicinal herbs, and to provide technical support for the quantitative description of the microscopic identification characteristics of Chinese herbal medicines. Methods: According to the characteristics of main morphological structures in the microscopic images of medicinal herbs, the digital image processing technology and the method of programming with MATLAB software were used to ingest and encode the microscopic images of each field of view of the cross-section of windproof medicinal herbs. The template matching method was used to gray the various fields of view. Degree images are registered and stitched, and the edge detection of the stitched image is performed by using Canny operator. The stitching image is binarized and the various microscopic feature objects are segmented and their sizes, shapes, and quantities are analyzed by basic morphological operations. The cluster analysis was used to characterize the distribution of tubing and catheters. The data of the four wind-proof herbs extracted from the regions were discriminated and analyzed. RESULTS: The main microscopic features of the cross-sections of the cork layer, phloem, xylem, tubing, and catheters of windproof herbs were segmented from the sample images; 20 parameters were extracted from the segmented images, and were selected by stepwise discriminant analysis. Twelve parameters were used to discriminate and analyze the differences in histomorphology of the samples of windproof medicinal materials from four producing areas, with a total accuracy rate of 91.4%. Conclusion: This method can accurately describe the subtle differences in the tissue structure of the cross-section of the windproof herbs from the four producing areas, which has certain reference significance for the microscopic identification of the wind-resistant medicinal materials from different habitats.