基于Lab颜色空间的自适应K均值彩色图像分割方法

Adaptive K-means Color Image Segmentation Method Based on Lab Color Space

为解决传统K均值算法在RGB空间处理彩色图像分割时,出现分割精度低、色彩表现差、初始聚类中心位置和数目难以确定等一系列问题,提出一种基于Lab颜色空间的自适应K均值彩色图像分割方法。首先将图像由默认的RGB空间转换到色彩表现更符合人眼机制的Lab空间;其次引入DBI指数作为聚类是否进一步分裂的判别依据,与改进的最大最小距离方法相结合,获得全局初始聚类中心点;最后运行k均值算法,进行彩色图像分割。实验结果显示,改进算法的平均运行时间较传统算法快6.316s,误差概率下降18.712%。该方法不仅解决了分割图色彩饱和度不足问题,并且获得了更快的分割速度和更准确的分割精度。

In order to solve a series of problems such as low segmentation accuracy,poor color performance,and difficulty in determining the location and number of initial cluster centers when the traditional K-means algorithm processes color image segmentation tasks in RGB space.An adaptive K-means color image segmentation method based on Lab color space is proposed:firstly,the image is converted from the default RGB space to the Lab space whose color performance is more in line with the human eye mechanism;secondly,the DBI index is introduced to determine whether the cluster is further split The discrimination basis is combined with the improved maximum and minimum distance method to obtain the global initial cluster center point;finally,the K-means algorithm is run for color image segmentation.The experimental results show that the average running time of the improved algorithm is 6.316s faster than the traditional algorithm,and the error probability is reduced by 18.712%.It not only solves the problem of insufficient color saturation of the segmentation map,but also obtains faster segmentation speed and more accurate segmentation accuracy.