基于大气散射模型与斜率变换的红外图像增强

Infrared image enhancement based on atmospheric scattering model and slope transformation

摘要:

为了获得效果更理想的红外图像,提出了结合大气散射模型与斜率变换的红外图像增强方法。根据像素本身的像素值及其邻域像素的均值,将图像划分为高亮区域与非高亮区域,分别用改进的暗通道先验和标准的暗通道先验估计其透射率；然后,选取透射率最小的原图像中的像素值,作为大气光估计值,进而根据大气散射模型进行图像恢复；最后,提出了一种非线性斜率变换函数,对恢复的红外图像进行变换,以增强图像的亮度和清晰度。实验数据表明,与最新提出的部分图像增强方法比较,本文方法处理后的图像在亮度、对比度和清晰度上的效果更佳,对应的信息熵和平均梯度分别提升2.08%和5.92%以上,因此本文方法能更有效地应用于实际的红外图像增强。

Abstract:

In order for more ideal infrared image, an infrared image enhancement method that combines atmospheric scattering model with slope transformation is proposed. The image is segmented into highlighted and non-highlighted areas based on the value of pixel itself and the mean of its neighborhood, for the two types of regions, their transmittances are estimated using the improved dark channel prior and the standard dark channel prior, respectively. And the pixel value in the original image with the lowest transmittance is selected as the estimated atmospheric light value, and then image restoration based on the atmospheric scattering model is performed. Finally, a nonlinear slope transformation function is proposed for the restored infrared image, with the aim to enhance its brightness and clarity. Experimental data show that compared to some latest image enhancement methods, proposed method performs better in terms of brightness, contrast, and clarity of the image, the corresponding information entropy and average gradient increase by 2.08% and 5.92%, respectively. Therefore, this method can be applied to actual infrared image enhancement more effectively.