图像分割中区域灰度重叠问题研究  

Study on the overlap of regional intensity ranges in image segmentation

作　　者：曹建农[1] 孙承志[2] 

机构地区：[1]长安大学地球科学与国土资源学院,陕西西安710064 [2]国家测绘局国土测绘司,北京100830

出　　处：《计算机应用》2006年第9期2077-2080,共4页journal of Computer Applications

基　　金：地理信息工程国家测绘局重点实验室基金资助项目(1469990324233);国家863计划资助项目(2001AA135081)

摘　　要：通过对图像不同区域灰度强度重叠问题以及图像平滑与滤波概念差异的分析,发现图像平滑对于无噪声污染图像的分割具有抑制次要目标、加剧区域边界渐变的作用,认为对图像进行平滑处理可以减小区域灰度重叠,产生过渡区域便于图像中特定目标的分割。实验表明图像灰度重叠是实现正确分割的障碍,图像平滑可以消除或减弱它的影响。This paper studied the overlap of regional intensity ranges in image segmentation, and probed into the concept difference between image smoothing and filter. It was found that image smoothing can were decrease minor objects and aggravate gradual change for segmentation. Therefore, it was thought that image smoothing could reduce the overlap of regional intensity ranges and bring out transition region for the segmentation of specific image objects. The experiments indicate that the overlap of regional intensity ranges in image is obstacle to correct image segmentation, and image smoothing can eliminate or weaken its impact.

关 键 词：图像平滑 图像滤波 图像分割 过渡区 灰度重叠 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]