基于灰度和局部熵迭代的红外目标分割算法

图像阈值分割在图像分析和图像识别中具有重要的意义,给出了一种以改进的最大类间方差法为基础的自适应阈值图像分割方法,同时利用分割后目标和背景区域的灰度信息和局部熵信息,设计了一个判断是否得到正确分割的准则,通过迭代循环,完成对图像的自动分割.实验结果表明,本文算法自适应性强,可以快速、准确、完整地分割出复杂背景图像中不同大小的红外目标.

Infrared Image Segmentation Algorithm Based on Iteration of Gray Level and Local Entropy

Image threshold segmentation is of important significance for image analysis and recognition. A adaptive method of automatic threshold selection based on the improved Otsu thresholding is given. Taking the gray level statistics and local entropy of the resultant target and background region as a criterion for discrimination, the segmentation whether is right or not shall be judged and the purpose of image segmentation shall automatically be carried out by using iteration. The result of experiment shows that the method has a good adaptive, and infrared targets with different dimension can be segmented out of complex background quickly, completely and rapidly.