摘要
阈值法是图像分割的一种重要方法,在图像处理与目标识别中广为应用.信息熵可以表征图像的灰度信息,并用以区分图像中的目标和背景.文中研究了最大熵法的分割效果、对数熵的运算时间,然后使用指数熵代替对数熵,并对二维最大熵法进行了改进,在结合大津法的同时,加入了4邻域外像素灰度的信息.实验结果表明本文所用方法可有效缩短计算时间、突出边缘特征、提高阈值自动选择的准确性和鲁棒性.
Thresholding is an important method of image segmentation and is applied widely to image processing and object recognition.Information entropy can characterize the grayscale information of image and also can distinguish between the objectives and background.After studying the segmentation effect of the maximum entropy method and the time required for computing the logarithmic entropy,two-dimensional maximum entropy method was improved in this paper.The method combined with the Otsu method.Simultaneously,it used exponential entropy instead of logarithmic entropy and joined the grayscale information of the pixels outside the four-neighbor-domain.Experiment results show that the improved method is effective to reduce the computing time and highlight the edges.Also,it can raise the accuracy and robustness of the automatic threshold selecting.
出处
《江苏科技大学学报(自然科学版)》
CAS
北大核心
2010年第5期485-488,共4页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
关键词
图像分割
阈值分割法
阈值选择
熵
image segmentation
threshold segmentation method
threshold selection
entropy