【图像增强】基于人工多重曝光融合AMEF实现图像去雾附matlab代码

1 简介

 Bad weather conditions can reduce visibility on images acquired outdoors, decreasing their visual quality. The image processing task concerned with the mitigation of this effect is known as image dehazing. In this paper we present a new image dehazing technique that can remove the visual degradation due to haze without relying on the inversion of a physical model of haze formation, but respecting its main underlying assumptions. Hence, the proposed technique avoids the need of estimating depth in the scene, as well as costly depth map refinement processes. To achieve this goal, the original hazy image is first artificially under-exposed by means of a sequence of gamma-correction operations. The resulting set of multiply-exposed images is merged into a haze-free result through a multi-scale Laplacian blending scheme. A detailed experimental evaluation is presented in terms of both qualitative and quantitative analysis. The obtained results indicate that the fusion of artificially under-exposed images can effectively remove the effect of haze, even in challenging situations where other current image dehazing techniques fail to produce good-quality results. An implementation of the technique is open-sourced for reproducibility

2 部分代码

function amef_demo% This set of matlab files implements the image dehazing method explained% in:% Galdran, A., "", 2018% which has been accepted publication in Elsevier's Signal Processing% Journal.% The fusion part relies heavily in the method presented in:%   "Exposure Fusion"%   Tom Mertens, Jan Kautz and Frank Van Reeth%   In Proceedings of Pacific Graphics 2007% If you find useful this method, please consider citing the original % paper. If you are only interested in the fusion part of the technique,% please consider citing Mertens' work.​clcclear all​% process a demo imagepath = '../demo_images/';% images shown in the paper% im_name = [path, 'road_input.png']; % Fig. 10% im_name = [path, '89.bmp']; % Fig. 7% im_name = [path, '91.jpg']; % Fig. 9% im_name = [path, '98.bmp']; % Fig. 1% im_name = [path, 'HazyDay.png']; % Fig. 8im_name = [path, 'horses.jpg']; % Fig. 6% im_name = [path, 'mumbai.jpg']; % Fig. 13% im_name = [path, 'landscape2-Tangkt.png']; % Fig. 11% other examples% im_name = [path, '2.jpg'];% im_name = [path, '51.jpg'];% im_name = [path, '88.jpg'];% im_name = [path, '99.png'];% im_name = [path, '100.png'];% im_name = [path, 'fc7.jpg'];% im_name = [path, 'IMG_4681_resize.jpg'];% im_name = [path, 'Bu_Wf_D_L_040.jpg'];% im_name = [path, '40.jpg'];% im_name = [path, 'beijing3.png'];​​​​​I_hazy = imread(im_name);% Increase clip_range to remove more haze - at the risk of overenhancementclip_range = 0.010;​ticamef_im = amef(im2double(I_hazy), clip_range);time = toc;​​[m, n, ~] = size(I_hazy);disp(['Resolution: ', num2str(m), ' x ', num2str(n)])disp(['Processing time: ', num2str(time)])​​figure(1),subplot(121); imshow(I_hazy);title('原图'),subplot(122), imshow(amef_im);title('增强图')​end

3 仿真结果

4 参考文献

[1]Galdran. Image dehazing by artificial multiple-exposure image fusion. ​

博主简介：擅长智能优化算法、神经网络预测、信号处理、元胞自动机、图像处理、路径规划、无人机等多种领域的Matlab仿真，相关matlab代码问题可私信交流。

部分理论引用网络文献，若有侵权联系博主删除。