在水下的環境中，由於影像受到水中之懸浮物質的影響，我們無法得到清晰的影像資訊。光在水中的傳遞過程中，會遇到光的吸收以及散射等問題，最主要還是因為散射導致影像霧化以及水中傳遞時波長的衰減程度不一而造成的色偏現象。水下影像在水中的霧化情況與在霧氣中拍攝的影像非常相似，都有著對比度下降以及色彩偏移的情形。未來若是要將除霧方法應用至監視系統上，除了效果之外，我們亦必須考慮到運算量的多寡。
最近在 [1]中提出了暗通道預設(Dark Channel Prior)的方法，可有效進行單張影像除霧，而有數個研究(例如 [2-7])已經開始將此方法用於水下影像的強化。在本篇論文中，我們以影像除霧之暗通道預設原理來改進水下影像除霧的方法，以及運用均值濾波器( Mean Filter)的技術來估計物體至相機距離的距離深度圖並估計水下背景光源，接著求出霧氣的透射平面(Transmission Map)，再使用快速導引濾波器(Fast Guided Filter)來修正邊緣，減緩光暈的產生，最後就能進行還原，並得到無霧影像。處理完水下影像的霧氣之後，接下來再針對水下影像的其他情形進行修正。
基於暗通道預設的水下影像還原方法中，透射平面的估計是重要的步驟。為了獲得良好的視覺效果，通常採用導引濾波器來修正邊緣的部分。本論文中，我們首先以導引濾波器為基礎，並進行簡化使其適用於水下影像的修復，以降低計算的複雜度並保持良好的視覺效果。接著，本論文提出一個基於簡化導引濾波器的低成本架構，以實現即時Full HD(1920  1080)的水下影像修正。更進一步，本論文以提出之簡化導引濾波器為基礎，再提出一簡化快速導引濾波器架構，以影像縮放之概念，使得速度大幅提升，並且進一步降低記憶體的使用量以降低面積，並且在視覺上也能維持優良的效果。

In an underwater environment, we can’t get a clear image because the image is affected by suspended solids. Moreover, the absorption and scattering of light and other issues will occur in the process of underwater transmission. Therefore, the main problems of underwater images are that the scattering phenomenon causes the image to become hazy and the varying degree of attenuation for the light with different wavelength in the water leads to the color distortion. Underwater image is very similar to common hazy image, both of them have the problems of low contrast and color distortion. In the future, if we want to apply the dehazing method to the surveillance system, we must also consider the computational complexity.
Recently, a method of Dark Channel Prior has been proposed in [1] to effectively perform single image dehazing, and several studies (e.g., [2-7]) have adopted this method for underwater image enhancement. In this thesis, we improve the underwater image dehazing method based on the Dark Channel Prior. Furthermore, the mean filter is employed to estimate the distance from the object to the camera and the atmospheric light in the water. And then, the transmission map of the hazy image is found, and the edge is corrected by using the simplified fast guided filter to revise the edge. Finally, we can obtain the image without haze. After removing the haze of the underwater image, the next step is correcting the color distortion.
In the adopted underwater image restoration method based on Dark Channel Prior, the estimation of the transmission map is an important step. In order to obtain a good visual quality, the guided filter is usually used to correct the edge of the part. In this thesis, a simplified guided filter is proposed and applied to the repair of underwater images to reduce the computational complexity and maintain a good visual quality. In addition, a low-cost architecture of the proposed simplified guided filter is designed to achieve the real-time underwater image correction for the size of Full HD (1920  1080). Furthermore, based on the proposed simplified guided filter, we propose a simplified fast guided filter architecture, which significantly improves the speed of guided filtering through image scaling and further reduces the memory usage and area. After these simplifications, it is still able to maintain excellent visual quality.

致謝 ii
論文提要 iii
摘要 iv
Abstract v
目錄 vii
圖目錄 ix
表目錄 xii
第一章 序論 1
1.1 研究動機 1
1.2 論文大綱 3
第二章 研究背景 4
2.1 大氣散射原理 4
2.2 基於暗通道預設之除霧演算法 6
2.2.1 暗通道預設 6
2.2.2 估計傳遞率 8
2.2.3 估計大氣光值 9
2.2.4 無霧影像還原 10
2.3 邊緣修正方法 11
2.3.1 導引濾波器 12
2.3.2 快速導引濾波器 14
2.4 水下影像強化方法 16
第三章 研究方法 19
3.1 運算流程 19
3.2 計算水下之大氣光值 20
3.3 估計傳遞率 21
3.4 傳遞率修正 23
3.4.1 簡化導引濾波器 23
3.4.2 簡化快速導引濾波器 26
3.5 水下影像強化 30
3.5.1 水下影像還原 30
3.5.2 色偏修正 31
第四章 提出的水下影像強化硬體架構 33
4.1 硬體架構 33
4.2 傳遞率估計器 35
4.3 灰階轉換器 36
4.4 邊緣修正硬體架構 37
4.4.1 Mean Filter架構 37
4.4.2 簡化導引濾波器架構 39
4.4.3 簡化快速導引濾波器架構 41
4.5 還原模組 42
4.5.1 特殊函數運算單元架構 42
4.5.2 定點數−浮點數格式轉換器 44
4.5.3 浮點數−定點數格式轉換器 45
第五章 實驗結果 46
5.1 實驗步驟與方法 46
5.2 硬體與軟體實作效果比較 47
5.3 水下影像強化效果比較 58
5.4 硬體分析 63
第六章 結論與未來研究方向 65
6.1 結論 65
6.2 未來研究方向 65
參考文獻 66

[1] K. He, J. Sun, and X. Tang, “Single image haze removal using dark channel prior,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 33, no. 12, pp. 2341-2353, Dec. 2011.
[2] A. Galdran, D. Pardo, A. Picón, and A. Alvarez-Gila, “Automatic red-channel underwater image restoration,” Journal of Visual Communication and Image Representation, vol. 26, pp. 132-145, Jan. 2015.
[3] H. Y. Yang, P. Y. Chen, C. C. Huang, Y. Z. Zhuang, and Y. H. Shiau, “Low complexity underwater image enhancement based on dark channel prior,” Innovations in Bio-inspired Computing and Applications (IBICA), 2011 Second International Conference on IEEE, pp. 17-20, Dec. 2011.
[4] S. J. Horng, P. J. Liu, and J. S. Lin, “Improving the Contrast Enhancement of Oceanic Images Using Modified Dark Channel Prior,” in Proc. 2016 International Symposium on Computer, Consumer and Control, pp. 801-804, July, 2016.
[5] S. Borkar, and S. V. Bonde, “Underwater image restoration using single color channel prior,” in Proc. 2016 International Conference on Signal and Information Processing (IConSIP), pp. 1-4, Oct. 2016.
[6] C. Li, J. Quo, Y. Pang, S. Chen, and J. Wang, “Single underwater image restoration by blue-green channels dehazing and red channel correction,” in Proc. 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1731-1735. Mar. 2016.
[7] C. Y. Cheng, C. C. Sung, and H. H. Chang, “Underwater image restoration by red-dark channel prior and point spread function deconvolution,” in Proc. 2015 IEEE International Conference on Signal and Image Processing Applications (ICSIPA), pp. 110-115, Oct. 2015.
[8] K. He, J. Sun, and X. Tang. “Guided image filtering,” IEEE transactions on Pattern Analysis and Machine Intelligence vol. 35, no. 6, pp. 1397-1409, June. 2013.
[9] C. C. Kao, J. H. Lai, and S. Y. Chien, “VLSI architecture design of guided filter for 30 frames/s full-HD video.” IEEE Transactions on Circuits and Systems for Video Technology, vol. 24, no. 3, pp. 513-524, Mar. 2014.
[10] K. He, and J. Sun, “Fast guided filter,” ArXiv preprint ArXiv:1505.00996, 2015.
[11] 影像處理實驗室http://image.cse.nsysu.edu.tw/index1.html
[12] C. Kuo, “Underwater Image Dehazing and Enhancement by Compensating Light Absorption Loss and Multi-exposure Fusion,” 國立中山大學, 碩士論文, 2017.
[13] C. Ttofis, C. Kyrkou, and T. Theocharides, “A Low-Cost Real-Time Embedded Stereo Vision System for Accurate Disparity Estimation Based on Guided Image Filtering,” IEEE Transactions on Computers, vol. 65, no. 9, pp. 2678-2693, Sept. 2016.
[14] H. Lu, Y. Li, and S. Serikawa, “Underwater image enhancement using guided trigonometric bilateral filter and fast automatic color correction,” Image Processing (ICIP), 2013 20th IEEE International Conference on. IEEE, pp. 3412-3416, Sept. 2013.
[15] C. Ancuti, C. O. Ancuti, and T. Haber, “Enhancing underwater images and videos by fusion,” Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on. IEEE, pp. 81-88, June. 2012.
[16] D. Park, D. K. Han, and H. Ko, “Enhancing underwater color images of diving mask mounted digital camera via non-local means denoising,” Consumer Electronics (ICCE), 2016 IEEE International Conference on. IEEE, pp. 441-442, Jan. 2016.
[17] L. W. Hsu, “A Variable-precision Architecture for Special Function and Floating-point Multiply-add-fused Operation,” 國立中山大學, 碩士論文, 2015.
[18] J. Y. Li, “Design, Implementation, and Analysis of a Multi-Precision Rasterizer for 3D Graphics Application,” 國立中山大學, 碩士論文, 2016.
[19] W. Zou, X. Wang, K. Li, and Z. Xu, “Self-tuning underwater image fusion method based on dark channel prior,” Robotics and Biomimetics (ROBIO), 2016 IEEE International Conference on. IEEE, pp. 788-793, Dec. 2016.
[20] S. Bazeille, I. Quidu, L. Jaulin, and J. P. Malkasse, “Automatic underwater image pre-processing,” CMM'06, Oct. 2006.
[21] N. Carlevaris-Bianco, A. Mohan, and R. M. Eustice, “Initial results in underwater single image dehazing,” OCEANS 2010. IEEE, pp. 1-8, Sept. 2010.
[22] J. Y. Chiang, and Y. C. Chen, “Underwater image enhancement by wavelength compensation and dehazing,” IEEE Transactions on Image Processing, vol. 21, no. 4, pp. 1756-1769, April 2012.
[23] S. Serikawa, and H. Lu, “Underwater image dehazing using joint trilateral filter,” Computers & Electrical Engineering, vol. 40, no. 1, pp. 41-50, Jan. 2014.