结合纹理特征改进的GBIS图像分割方法

摘要

针对GBIS(efficient graph-based image segmentation)方法在分割含有较丰富纹理信息的图像时,分割效果不理想的问题,在L*a*b*彩色空间下,结合图像的纹理特征,提出了一种改进GBIS图像分割方法,记为IG-BIS(improved efficient graph-based image segmentation).该方法首先将图像由RGB空间转换到L*a*b*颜色空间；接着,结合L*a*b*彩色空间,对GBIS方法中的权值函数作了改进,引入了一个常数s,用于控制相邻像素之间颜色的差异程度；然后,用熵的方法来获取L*a*b*彩色图像的纹理特征；最后,结合图像的纹理信息,改变了GBIS方法中的区域合并条件,得到最终的分割结果.实验证明,与原算法相比,该方法在分割精度与分割质量上有了很大程度的提高.IGBIS有效地抑制了彩色图像在分割中存在的过分割现象,并适合于含有丰富纹理的彩色图像.%According to the problem of the GBIS method that the result wasn't ideal when the image contained rich texture information,combined with the texture features of the image,the paper proposed an improved GBIS image segmentation algorithm(IGBIS)in L* a* b * color space.Firstly,the method converted the image from RGB color space to L* a* b* color space.Secondly,the weight function in GBIS method was improved by a constant s that it introduced to control the color difference degree.Thirdly,it obtained texture features of the image in L* a* b* color space by using entropy method.Finally,combined with the texture features of the image,it changed the region merging condition in GBIS method,got the final segmentation result.Experiments show that,compared to source GBIS,the method is gready improved in the accuracy and quality of segmentation.In color image segmentation,IGBIS method effectively inhibites over-segmentation phenomenon,and is suitable for the image contained rich texture information.