像素级自适应融合的夜间图像增强

Night image enhancement based on pixel level adaptive image fusion

针对夜间图像整体亮度低、图像细节多集中在较低灰度级范围内的特点,本文提出了一种像素级自适应融合的夜间图像增强方法。首先,通过高通滤波的方式滤除图像的大部分暗背景信息,以整体增强图像各区域灰度细节;然后,将滤波后的图像与原始图像进行像素级自适应加权融合,其中每个像素点对应的两个权值取决于滤波图像内对应像素值与目标模糊集、背景模糊集的接近情况:越接近目标模糊集,滤波图像对应像素权值增大,同时原始图像对应像素权值减小;越接近背景模糊集,则相反,以此来进一步增强图像的全局对比度,增强视觉效果。实验结果表明:增强后的夜间图像在灰度细节和全局对比度方面都得到了显著提高。对于320×256分辨率的灰度图像,在CPU主频为3.0 GHz的硬件环境下,处理时间小于10 ms。采用本文方法处理的夜间退化图像的成像质量得到了明显改善,且其良好的实时性能够更好地满足工程应用需求。

Aiming at the characteristics of night images,such as low brightness and most of the image details concentrating in the low gray range,a night image enhancement method based on pixel level adaptive fusion is proposed.Firstly,a high pass filter is used to filter most of the dark background information of the image,so as to enhance the gray details of each region of the image.Then,the filtered image is fused with the original image in the way of pixel level weighting adaptively.The two weights of each pixel,in filtered image and original image respectively,depend on two distances:one is the distance between the pixel value and the target fuzzy set,and the other is the distance between the pixel value and the background fuzzy set.If the pixel value is closer to the target fuzzy set,the weight of the filter image is increased and the weight of the original image is reduced,correspondingly.If the pixel value is closer to the background fuzzy set,the case is the opposite.In this way,the global contrast of the image is further enhanced and the visual effect is improved.Experimental results indicate that the night image enhanced has been improved significantly in both gray detail and global contras.The processing time is less than 10 ms in the hardware environment of CPU with the main frequency of 3.0 GHz for 320×256 resolution gray scale images.The quality of the degraded night image has been improved greatly,and its good real-time performance can meet the needs of the engineering application better.