论文部分内容阅读
本文以水泥窑生料球为背景，分析了球形物料图像在灰度空间和梯度空间的特征，对一些经典算法进行了颗粒分割实验，理论分析和实验结果表明，利用局部特征值进行总体像素归属判决将使颗粒图像的边界不连续，或边界描述失真。为此提出的一种由局部特征值直接提取球形物料边界的图象分割方法。实验结果表明，该算法分割的图像边界封闭性好，边界描述真实，适合球形物料图像的分割。
In this paper, the raw material ball of cement kiln is taken as the background, the characteristics of spherical material images in gray space and gradient space are analyzed, and the particle segmentation experiments of some classical algorithms are carried out. The theoretical analysis and experimental results show that using local eigenvalues The verdict will make the boundaries of the particle image discontinuous or the boundary description distorted. To this end, an image segmentation method of extracting the boundary of spherical material directly from local eigenvalues is proposed. The experimental results show that the proposed algorithm has the advantages of good image segmentation and good boundary description, and is suitable for the segmentation of spherical material images.