基于高斯多峰拟合与直方图规定化的红外图像增强  

Infrared Image Enhancement Based on Gaussian Multimodal Fitting and Histogram Normalization

作　　者：闫娜[1] 崔灿[1] 王晓曼[1] YAN Na;CUI Can;WANG Xiaoman(School of Electronics and Information Engineering,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学电子信息工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2018年第3期115-119,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

摘　　要：为提高红外图像清晰度,提出一种基于高斯多峰拟合和直方图规定化的红外图像增强算法。首先对图像直方图进行平滑处理,通过求导获得直方图波峰数目,对直方图进行高斯多峰拟合,并通过BML映射规则获得规定化图像。为补偿丢失的弱边缘信息,采用四方向Sobel算子获得原图的梯度,利用梯度对规定化图像进行锐化以增强图像轮廓细节。实验表明,经过该算法处理的红外图像,可识别度和层次感都明显优于传统算法,使红外图像视觉效果得到很大的改善。In order to improve the sharpness of infrared image, an infrared image enhancement algorithm based on Gaussian multimodal fitting and histogram normalization is proposed. Firstly,the histogram of the image is smoothed,and the number of histogram peaks is obtained by derivation. The histogram is Gaussian multimodal fitting,and the normalized image is obtained by BML mapping rules. In order to compensate for the missing weak edge information,the four-direction Sobel operator is used to obtain the gradient of the original image, and the normalized image is sharpened by the gradient to enhance the detail of the image contour. Experiments show that the infrared image,the recognition degree and the layering effect are obviously superior to the traditional algorithm,and the visual effect of the infrared image is greatly improved.

关 键 词：红外图像增强 多峰拟合 直方图规定化 映射规则 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]