基于灰度聚类算法的红外图像增强研究

根据红外灰度图像的特点，提出了一种基于K-均值聚类的图像增强的新算法。该算法首先根据具体图像确定K值，其次对红外图像的辐射温度数据进行统计学习，把不同温度值按升序排列，然后按等差原则选取温度值作为初始聚类中心，再依据初始聚类中心采用K-均值聚类算法对温度进行聚类，最后由聚类结果对图像进行自适应增强。通过对红外灰度图像进行实验，得到了满意的结果: 对比直方图均衡，具有更丰富的图像细节信息和层次感，视觉效果更好。

Infrared image enhancement based on gray clustering algorithm

A new algorithm method of image enhancement based on K-means clustering is presented, according to the characteristics of infrared gray-image. In this method, firstly, the value of K is determined with the specific image, then the statistic analysis is carried out for the radiation temperature data of the infrared image to sort the data in ascending order, finally, the data of an arithmetical progression are selected as initial clustering centers and the temperature data are clustered by K-means algorithm with the generated clustering centers. At last, the self-adaptive enhancement is completed for the image with the clustering result. The satisfactory result is achieved in the experiment with gray infrared images. Compared with the histogram equalization method, the simulation result indicates that this method could further improve the visual effect, and provide detailed information and better gradation.