A Novel Robust Algorithm for Image Segmentation

One-dimensional threshold segmentation algorithm has poor adaptability and poor noise immunity, while two-dimensional threshold segmentation has high computational complexity. To solve these problems, we propose a novel threshold segmentation algorithm for image segmentation. The whole method contains three steps: constructing one-dimensional histogram by using each pixel´s gray value and its surrounding area´s mean gray value, by integrating three classic segmentation algorithms: minimum error, maximum entropy and Otsu, we construct a new threshold selection algorithm, making use of the threshold to instruct segmentation. The experimental results show that our algorithm has strong adaptability, stable and robust noise immunity when comparing to one-dimensional threshold segmentation algorithm and a wider range of noise adaptability, lower computation complexity when comparing to two-dimensional Otsu algorithm.