基于图像分割和平台直方图均衡的红外图像增强算法

针对原始红外图像模糊、对比度低的特点,提出了一种基于图像分割的平台直方图均衡算法。首先采取实时阈值提取方法选取阈值,将图像目标灰度与背景灰度分开;然后利用对数、指数结合的非线性变换抑制背景区域,扩展目标区域的动态范围;最后利用改进的双平台直方图均衡算法处理非线性变换后的灰度图像。实验结果表明:该方法有效地增强了原始红外图像,提高图像对比度的同时很好地保留了图像的细节和边缘信息,且避免了过亮现象,视觉效果得到改善。

Infrared Image Enhancement Algorithm Based on Image Segmentation and Platform Histogram Equalization

For the fuzzy and low contrast of original infrared images, an algorithm based on image segmentation and platform histogram equalization was proposed. Firstly, we took real-time threshold extraction method to choose threshold that separated the target gray and the background gray of image. Secondly, we used the nonlinear transform which was the combination of logarithm and index to inhibit background region and expand dynamic range of the target area. Finally, the improved double platform histogram equalization algorithm was used to process the nonlinear transformed grayscale image. Experiment results show that the method can enhance the original infrared image effectively and improve the contrast and reserve the details and edges of the image well, moreover, the method avoid over-enhancement and improve visual effect.