摘要
比较、评价了各种经典的图像二值化方法对ICT图像二值化处理的效果。综合分析及实践发现,实验法、大津法不适用于对ICT图像进行二值化分割;而迭代法、边缘检测阈值法、加权边缘检测阈值法对ICT图像二值化分割效果不好。直方图法处理的效果稍好,但它仍将非常接近背景灰度的目标物部分错误划分为背景。提出“背景法”、直方图法与背景法结合的改进方法,可对ICT断层图像进行较理想的二值化分割。
The biggest difficulty is to automatically produce an optimal threshold value used in image segmentation. In this paper, it compares and appraises the effects of several different classical image binarization approaches. According to the synthesized analysis and practice, it finds that the experimental method and the Dajin method are unsuitable to segment ICT images, and the other classical image binarization approaches cannot segment ICT images properly even if the result of the histogram method is a little better. However, the approaches given by the author can segment ICT images relatively ideally.
出处
《制造技术与机床》
CSCD
北大核心
2006年第9期54-57,共4页
Manufacturing Technology & Machine Tool
基金
重庆市科学技术委员会的科技攻关计划项目
关键词
ICT图像二值化
图像分割
阈值选择与确定
ICT Image
Binarization
Image Segmentation
Selection and Definition of Optimal Threshold Value