基於暗通道先驗之適應性快速影像除霧方法

Adaptive Fast Image Dehazing Algorithms Based on Dark Channel Prior

林育聖(Yu-Sheng Lin)

指導教授 ： 謝政勳

朝陽科技大學/資訊學院/資訊工程系/碩士(2014年)

摘要

一般數位影像容易因為環境變化而影響輸出的影像的品質。例如當受到霧氣及霾的影響常會使影像的色彩飽和度及對比度降低。因此需要透過使用影像除霧方法來強化影像的品質。儘管現今已有許多的影像除霧方法被提出，然而大多因為運算複雜而需要較長的處理時間，因此我們提出一個基於暗通道先驗之適應性快速影像除霧方法，其中使用1×1視窗來計算暗通道並以暗通道圖來估計大氣環境光，以此解決光暈的問題、降低計算複雜度及過度曝光的問題。透過多個範例驗證本論文所提的除霧方法，與其他學者的除霧演算法比較，證明我們的方法具有較佳的效率，並且能得到令人滿意的除霧影像。

關鍵字

高斯濾波 ； 暗通道先驗 ； 影像除霧 ； 影像融合

並列摘要

The quality of digital images is easily affected by imaging conditions. Haze is one of adversarial conditions to degrade the image quality, such as contrast, visibility and color saturation. Many approaches have been proposed to deal with the hazy condition. However, most of them suffer from high computational complexity. In this thesis, several adaptive fast image dehazing algorithms based on the dark channel prior are presented. In the proposed dehazing algorithms, dark channel map is found through 1×1 minimum filter and then used to estimate the atmospheric light and transmission map. By this doing, the computation complexity is reduced and over-exposure problem generally happened in the dark channel based algorithms is avoided. Simulation results of several examples indicate that the proposed dehazing algorithms are generally able to obtain satisfactory dehazed images and are more efficient than the compared dehazing algorithms.

並列關鍵字

Gaussian filtering ； dark channel prior ； image dehazing ； image fusion

參考文獻

[9] Jin-Hwan Kim, Won-Dong Jang, Jae-Young Sim, Chang-Su Kim, “Optimized Contrast Enhancement for Real-time Image and Video Dehazing,” Journal of Visual Communication and Image Representation, Vol. 24, Iss. 3, pp. 410-425, February, 2013.

[1] R. Tan, “Visibility in Bad Weather from a Single Image,” Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Washington, DC, pp. 2347-2354, 2008.

[2] R. Fattal, “Single image dehazing,” ACM Transactions on Graphics, Vol. 27, No. 3, pp. 721-729, 2008.

[3] Kaiming He, Jian Sun, and Xiaoou Tang “Single Image Haze Removal Using Dark Channel Prior,” IEEE Transactions on Pattern and Analysis and Machine, Vol. 33, Issue 12, pp. 2341–2353, Dec., 2011.

[5] Kristofor B. Gibson, Dung T. Võ, Truong Q. Nguyen, “An Investigation of Dehazing Effects on Image and Video Coding”, IEEE Transactions on Image Processing, Vol. 21, No. 2, pp. 662-673, 2012.