Histogram equalization (HE)-based technology has been widely applied in infrared image contrast enhancement due to its effectiveness and simple implementation. However, HE and its variations considered the accumulation of pixels in different gray values, thus ordinarily result in artifact effects, over-enhancement and noise amplification, especially in the uniformity region. In this paper, we redefine and formulate a new HE technology to overcome the shortcomings of traditional HE technology. 2D difference information between two adjacent pixels is introduced for infrared image histogram calculation and its calculation is achieved by a reasonable difference threshold. With the purpose of adaptability to different scenes, the display range of output image is controlled by 2D difference information. To preserve detail edges, we apply adaptive plateau HE on 2D difference information-related histogram. Experiments and results show that the proposed algorithm has better scene adaptability and outperforms other compared algorithms by enhancing the contrast without introducing over-enhancement effect.

基于直方图均衡（HE）的技术由于其有效性和简单的实现方式已广泛应用于红外图像对比度增强。然而，HE及其变化考虑了不同灰度值的像素的累积，因此通常导致伪像效应，过度增强和噪声放大，尤其是在均匀性区域中。在本文中，我们重新定义和制定了一种新的HE技术，以克服传统HE技术的缺点。引入两个相邻像素之间的二维差异信息以进行红外图像直方图计算，并通过合理的差异阈值来实现其计算。为了适应不同场景，输出图像的显示范围由2D差异信息控制。为了保留细节边缘，我们将自适应高原HE应用于2D差异信息相关直方图。实验和结果表明，该算法在不引入过度增强效果的前提下，增强了对比度，具有较好的场景适应性，并且优于其他算法。