Abstract: The infrared image has large noise and low contrast characteristic. The infrared image enhancement is one of core problems of infrared detection,  recognition and tracking applications. Histogram equalization plays an important role in infrared image enhancement techniques, but large defect exists, for example the loss of detail information. This paper proposes an infrared image enhancement method based on subsection histogram with nonlinear transformation. The pixel luminance from the dark region is improved by nonlinear transformation. The histogram is divided into two sections based on foreground and background regional characteristics, and the two histograms are equalized to enhance the contrast of the foreground and background. The experiment proves that this method can effectively improve the brightness of the image; extend the gray range from the target region and enhance the detail in the foreground.

Key words: infrared image, image enhancement, nonlinear transfer, bi-histogram equalization

摘要： 红外图像具有噪声大、对比度低等特点，红外图像增强是红外探测、识别和跟踪应用中的核心问题之一。在红外图像增强技术中，直方图均衡方法简单、有效，但存在细节信息损失较大的缺陷。提出一种对红外图像采用非线性变换分段直方图的增强方法，该方法对红外图像进行非线性变换，提高较暗区域的像素亮度，根据前背景区域特征将直方图分成两段，进行双直方图均衡化处理，对前景和背景分别进行图像的增强。经过实验验证，该算法能有效提高图像亮度，扩大目标区的灰度范围，增强前景图像的细节部分。

关键词: 红外图像, 图像增强, 非线性变换, 双直方图均衡