摘要： 提出基于万有信息定律的图像阈值分割方法。针对熵阈值法仅利用图像灰度概率信息，导致它对有些图像的分割无效，该文从万有引力定律中得到启发，提出信息场中万有信息定律，将其用于图像分割的最佳阈值选取。实验结果表明，该方法是可行的，且对有些图像的分割效果要好于传统的Kapur熵方法。

关键词: 图像分割, 阈值法, 信息熵, 信息能量, 万有信息定律

Abstract: This paper proposes a new image thresholding segmentation method based on universal information law. Aiming at the disadvantage of entropy based thresholding method only making use of image gray probability information, and leading to be of no effect in segmentting some image, this paper presents universal information law in the information field in view of enlightenment from universal gravitation force law, and it is used to select optimal threshold of image segmentation. Experimental results show that the segmentation method of this paper is feasible, and segmentation quality of some image is much better than that of classical Kapur’s method based on Shannon’s entropy.

Key words: image segmentation, thresholding method, information entropy, information energy, universal information law

中图分类号: 

• TP391.41