基于改进K-means聚类的木材缺陷彩色图像分割算法研究

为实现木材缺陷图像的准确分割，研究了在RGB色彩空间使用K-means聚类算法进行木材缺陷图像的分割方法，针对K—means聚类算法图像分割的不足之处，并提出了一种改进方案。在改进算法中，加入对图像像素点的邻域处理，根据邻域特性适时调整原图像，充分利用图像像素的区域特征，以此来抑制一些局部噪声。实验结果表明，改进算法不仅能够滤除较多的干扰信息，而且能更好的拟合分割边界，改善图像的分割效果。

Study of Wood Defect Color Image Segmentation Algorithms Based on Improved K-means Clustering

In order to achieve the accurate segmentation of wood defect images in the RGB color space, we studied the application of K-means clustering algorithms on wood defect image segmentation. Aiming to the deficiencies of K-means clustering algorithm, we came up with an improved solution. We added further processing of the pixel neighborhood in the improved solution. Through adjusting the original image according to the neighborhood characteristics, we made full use of the regional characteristics of the image pixels in order to suppress the local noises. Experimental results showed that the improved algorithm can not only filter out the interference information, but also better fit the dividing boundary. It can improve the segmentation of the image.