基于数学形态学的植物叶片图像分割方法研究

摘要

Conventional edge operators are linear filtering method. Mathematical morphology, a methodology of nonlinear filters, can maintain the basic shape of image characteristics. It has been widely used in many fields of image processing. Only using edge detection operators or morphological processing could not obtain ideal segmentation effect. According to the results of inspection edge operator, for morphological filtering, mathematics morphological operation, then the different structure elements of plant leaves were selected repeatedly for testing edge information. The quantitative analysis but not the traditional method of experiments results showed that the edge detection was an effective method based on the division of the morphology.%传统的边缘检测算子均属于线性滤波方法.数学形态学是一种非线性滤波方法,能够保持图像的基本形状特征,在图像处理各个领域广泛应用.仅用边缘检测算子或者数学形态学处理都不能得到理想的分割效果.根据边缘算子检测的结果,进行形态学滤波、数学形态学运算,采取不同的结构元素反复实验对植物叶片边缘信息进行提取.不用传统目测方法而是通过定量的分析实验结果,表明了边缘检测基础上进行形态学处理的分割方法的有效性.