摘要
模糊理论对处理模糊不确定性的事件具有得天独厚的优势,而阈值分割算法是一种简单有效的算法,因此基于模糊理论的阈值分割在图像分割领域中得到广泛的应用。但是多数的阈值分割算法对于灰度图像的处理往往仅利用图像的灰度水平值而未考虑像素的空间邻域信息,大大地增加了灰度图像的目标和背景的误分比率。针对此问题提出一种引入空间信息的加权模糊阈值分割算法,在考虑像素的灰度水平值的同时,将像素与其邻域像素的空间关系作为权重来共同对该像素作用,设计出新的模糊分割函数。实验证明该算法具有较好的稳定性和鲁棒性,对多数图像的梯度的灰度直方图接近单峰的图像和灰度峰值分布距离较远、两峰数量悬殊的双峰图像,均能得到很好的分割效果。
Fuzzy theory has unique predominance in dealing with the events with fuzzy and uncertainties,and the threshold segmentation method is a simple and effective algorithm,therefore the fuzzy theory-based threshold segmentation has been widely applied in image segmentation field.However,when processing gray images,most of the threshold segmentation algorithms usually just utilise the gray level of the image but does not take into consideration the spatial neighbourhood information of pixels,this greatly increases the false rate of segmenting teh target from the background of gray image.In response to the problem addressed above,in this article we present a weighted fuzzy threshold segmentation algorithm with the spatial information introduced,while considering the gray level of pixels,spatial relationship of the pixel and its neighbouring pixel will be taken as the weight to jointly act on the pixel,thus a new fuzzy segmentation function is designed.Experimental result shows that this algorithm has good stability and robustness.It can achieve pretty good segmentation effects on most of the images,either the images with close-to-unimodal gray histogram of gradient or the twin-peak images with farther distribution distance of gray peak and in great disparity of number of each peak.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第3期274-277,326,共5页
Computer Applications and Software
关键词
模糊阈值分割
空间信息
图像分割
Fuzzy threshold segmentation Spatial information Image segmentation
