An Image Thresholding Method Based on Standard Deviation

Zuoyong Li, Chuancai Liu

Abstract: Classic statistical thresholding methods are not suitable for segmenting infrared images and fail to achieve ideal results. In this paper, a novel statistical thresholding method based on standard deviation is proposed for solving this problem. The algorithm utilizes standard deviations of two thresholded classes to define a new objective function and determines the optimal threshold by minimizing the function. The proposed algorithm was compared with two classic statistical thresholding methods on a variety of infrared images as well as general real world images, and the experimental results show the effectiveness of the algorithm.