【摘要】 在通过引导滤波方法对红外图像进行增强处理的过程中，图像显著区域与背景的对比度往往并没有得到有效增强，细节部分没有得到有效优化。针对此问题，提出了一种基于多尺度引导滤波的红外图像视觉显著增强算法。以引导滤波为基础将红外图像划分为基本层和细节层图像，基本层和细节层图像分别采用局部显著图权重融合与自适应Gamma校正方法进行增强处理；在此基础上，对红外原图、增强后的基本层与细节层图像按照权重因子进行特征融合，获得增强后的红外图像。实验结果表明：在三种不同场景下，所提算法的信息熵、峰值信噪比与图像标准差与其他算法相比均有明显提升。该算法更能突出图像中人眼所关注的重要细节部分，有效降低了噪声干扰，所得到增强后的红外图像更加清晰。更多还原

【Abstract】 In the process of infrared image enhancement by guided filtering method,the contrast between the prominent area and the background of the image is often not effectively enhanced,and the details are not effectively optimized. To solve this problem,a vision enhancement algorithm based on multi-scale guided filtering is proposed. Based on guided filtering,infrared images are divided into basic layer and detail layer images,which are enhanced by local saliency image weight fusion and adaptive Gamma correction method,respectively. On this basis,the infrared original image,the enhanced basic layer image and the detail layer image are fused according to the weight factor to get the final enhanced infrared image. Experimental results show that the information entropy,PSNR and image standard deviation of the proposed algorithm are significantly improved compared with other algorithms in three different scenarios. This algorithm can highlight the important details that human eyes pay attention to in the image,effectively reduce the noise interference,and the enhanced infrared image is clearer.更多还原