文中主要讨论了如何能将叶类中药显微细胞图快速有效的提取与识别，通过对细胞图深入的分析，对图像进行了形态学预处理，采用了中值滤波和边缘提取进行图像识别。针对标准的中值滤波方法存在的不足，结合融合的思想提出了改进的中值滤波算法。较标准的中值滤波具有更优良的滤波性能，更高的信噪比，同时拥有较低的漏检率和误检率。将滤波后的图像再进行边缘提取，更好地保存原图像的细节和边缘，对细胞进行了有效的识别，提取了细胞的特征，为后续细胞的参数计算奠定了基础。该算法容易实现、速度快，适合作为叶类中药显微图像的边缘检测及细化。%It discusses how they can figure leafy medicine microscopic cell and recognition quickly and effectively,through the analysis of cell diagram,preprocess the image in morphology,and recognize the image by median filtering and edge extraction. For the shortcomings of standard median filtering,combined the idea of integration,propose an improved median filter algorithm which has better filtering per-formance,a higher ratio of signal to noise,and a lower missed rate and false detection rate than standard median filtering. Through edge extraction on the filtered image,can better preserve the original details and edges of image,cells are identified effectively,extracting the cell characteristics,laid the foundation for the subsequent calculation of cell parameters. The algorithm is easy to implement and has a fast speed,so it's suitable for edge detection and refinement of leafy medicine microscopic image.
展开▼
