臺灣博碩士論文加值系統

在當前的水下研究領域中，水下影像經常面臨顏色失真的問題，這主要是由於自然光在水中的吸收和散射，以及不同水域的水質差異所致。特別是紅光的吸收速度最快，使得水下影像中的紅色往往會顯得較暗甚至消失，而藍光和綠光的穿透能力較強，使得深水區域的影像通常呈現藍綠色調。這些現象加上懸浮顆粒的散射效應，進一步降低了影像的清晰度和對比度。
為了解決這些問題，我們提出了UCED水下影像色彩還原模型。首先，我們提出了Mish激勵多層感知器（Mish Activation Multilayer Perceptron，MAMP）。與一般的MLP相比，MAMP在反向傳播過程中能夠保持穩定的梯度流動，避免梯度消失或爆炸的問題，同時保留負區間的部分資訊，有助於神經網路學習更複雜的特徵。接著，我們提出了LAB-RRBP（LAB Loss with Red-Reward and Blue-Penalty）和LCH-CR（LCH Loss with Chroma Reward）損失函數。LAB-RRBP損失函數能夠突出水下影像中的紅色部分，抑制藍色偏移的問題，而LCH-CR損失函數則提高了影像的色彩飽和度。
此外，我們根據紅光衰減比例提出了一套水深分級方法，將水下影像依據紅光吸收的比例分為三個等級，並根據相應的模型進行顏色還原，以解決水下顏色增強在不同海域中效果不佳的問題。這些方法顯著提升了水下影像的視覺效果和細節展現，使我們能夠更真實地觀察水下世界，並對其他科學研究如海洋生物學、地質學、考古學及環境監測等領域產生深遠的影響。透過提高影像品質，我們能更清晰地觀察和分析水下結構、遺跡及生態環境，進一步推動這些領域的發展。

In the current field of underwater research, underwater images often encounter issues of color distortion, primarily due to the absorption and scattering of natural light in water, as well as variations in water quality across different bodies of water. Notably, red light is absorbed at the fastest rate, resulting in the darkening or even disappearance of red hues in underwater images. In contrast, blue and green light have stronger penetration capabilities, causing images in deeper water areas to predominantly exhibit blue-green tones. These phenomena, compounded by the scattering effects of suspended particles, further reduce the clarity and contrast of the images.
To address these challenges, we propose the UCED model for underwater image color restoration. First, we introduce the Mish Activation Multilayer Perceptron (MAMP). Compared to conventional MLPs, MAMP ensures stable gradient flow during the backpropagation process, mitigating issues such as gradient vanishing or explosion, while also retaining information in the negative region, which aids the neural network in learning more complex features. Furthermore, we propose two loss functions: LAB Loss with Red-Reward and Blue-Penalty (LAB-RRBP) and LCH Loss with Chroma Reward (LCH-CR). The LAB-RRBP loss function enhances the red components in underwater images and suppresses the issue of blue shift, while the LCH-CR loss function improves color saturation in the images.
Furthermore, we propose a depth grading method based on the attenuation ratio of red light, categorizing underwater images into three levels according to the degree of red light absorption. Corresponding color restoration models are then applied to address the issue of suboptimal color enhancement across different marine environments. These approaches significantly enhance the visual quality and detail presentation of underwater images, enabling a more accurate observation of the underwater world. This advancement has profound implications for other scientific fields, including marine biology, geology, archaeology, and environmental monitoring. By improving image quality, we can observe and analyze underwater structures, relics, and ecosystems with greater clarity, thereby further advancing these disciplines.

目錄

中文摘要 I
Abstract II
致謝 III
目錄 IV
圖目錄 VI
表目錄 VII
第一章、研究背景與動機 1
1.1研究背景 1
1.1.1機器學習歷史 1
1.1.2機器學習類別 2
1.1.3機器學習應用 3
1.2研究動機 4
1.2.1水下機器學習應用 4
1.2.2水下影像顏色失真原因 5
1.3研究問題 7
1.4研究重要性 8
1.5論文架構 9
第二章、文獻回顧 10
2.1基於非物理模型 10
2.2基於物理模型 11
2.3基於深度學習 12
第三章、研究方法 17
3.1資料集準備 17
3.2網路架構 19
3.3損失函數 21
3.3.1 LAB顏色空間損失函數（LossLAB_RRBP） 22
3.3.2 LCH顏色空間損失函數（LossLCH_CR） 23
3.4水深分級 24
第四章、實驗結果與討論 28
4.1實驗設備與開發環境 28
4.2水下影像評估指標 29
4.2.1峰值訊噪比（PSNR） 29
4.2.2結構相似性指標（SSIM） 29
4.2.3水下影像品質測量（UIQM） 30
4.2.4水下顏色影像品質評估（UCIQE） 30
4.3水下顏色增強方法比較 31
4.3.1顏色增強後影像效果比較 31
4.3.2評估指標比較 38
4.4消融實驗 39
4.4.1閾值比較（Tchroma、Tred和Tblue） 39
4.4.2水深分類比較 41
第五章、結論 44
參考文獻 45

[1]A. M. Turing, Computing machinery and intelligence. Springer, 2009.
[2]J. McCarthy, M. L. Minsky, N. Rochester, and C. E. Shannon, "A proposal for the dartmouth summer research project on artificial intelligence, august 31, 1955," AI magazine, vol. 27, no. 4, pp. 12-12, 2006.
[3]F. Rosenblatt, "The perceptron: a probabilistic model for information storage and organization in the brain," Psychological review, vol. 65, no. 6, p. 386, 1958.
[4]A. Chaudhari, A. Vibhute, S. Barde, and M. Yerne, "Virtual Eye: Object Recognition for Blind," in 2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS), 2024: IEEE, pp. 1-5.
[5]Y. Xi, W. Jia, Q. Miao, J. Feng, J. Ren, and H. Luo, "Detection-Driven Exposure-Correction Network for Nighttime Drone-View Object Detection," IEEE Transactions on Geoscience and Remote Sensing, 2024.
[6]N. Manasa, D. W. Praveenraj, and L. SR, "Predictive Analytics for Stock Market Trends using Machine Learning," in 2023 4th International Conference on Computation, Automation and Knowledge Management (ICCAKM), 2023: IEEE, pp. 1-8.
[7]H. Yin and N. K. Jha, "A health decision support system for disease diagnosis based on wearable medical sensors and machine learning ensembles," IEEE Transactions on Multi-Scale Computing Systems, vol. 3, no. 4, pp. 228-241, 2017.
[8]H. Öztürk, A. Özgür, and E. Ozkirimli, "DeepDTA: deep drug–target binding affinity prediction," Bioinformatics, vol. 34, no. 17, pp. i821-i829, 2018.
[9]J. Rutledge et al., "Intelligent shipwreck search using autonomous underwater vehicles," in 2018 IEEE International Conference on Robotics and Automation (ICRA), 2018: IEEE, pp. 6175-6182.
[10]L. Huang, Y. Wang, Q. Zhang, J. Han, W. Tan, and Z. Tian, "Machine learning for underwater acoustic communications," IEEE Wireless Communications, vol. 29, no. 3, pp. 102-108, 2022.
[11]T. Gao, J. Yan, X. Yang, C. Chen, and X. Guan, "Integrated Target Recognition and Grasp Control for the Underwater Vehicle–Manipulator System," IEEE Transactions on Industrial Electronics, 2024.
[12]D. Akkaynak and T. Treibitz, "A revised underwater image formation model," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 6723-6732.
[13]J.-c. Zhou, D.-h. Zhang, and W.-s. Zhang, "Classical and state-of-the-art approaches for underwater image defogging: a comprehensive survey," Frontiers of Information Technology & Electronic Engineering, vol. 21, no. 12, pp. 1745-1769, 2020.
[14]N. Justin Marshall, "Communication and camouflage with the same ‘bright’colours in reef fishes," Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, vol. 355, no. 1401, pp. 1243-1248, 2000.
[15]R. C. Duarte, M. Stevens, and A. A. Flores, "Shape, colour plasticity, and habitat use indicate morph-specific camouflage strategies in a marine shrimp," BMC Evolutionary Biology, vol. 16, pp. 1-15, 2016.
[16]L. Kaufman, "Feeding behavior and functional coloration of the Atlantic trumpetfish, Aulostomus maculatus," 1976.
[17]Z. Liu, Y. Zhuang, P. Jia, C. Wu, H. Xu, and Z. Liu, "A novel underwater image enhancement algorithm and an improved underwater biological detection pipeline," Journal of Marine Science and Engineering, vol. 10, no. 9, p. 1204, 2022.
[18]Y. Y. Schechner and N. Karpel, "Clear underwater vision," in Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004., 2004, vol. 1: IEEE, pp. I-I.
[19]R. Hummel, "Image enhancement by histogram transformation," Unknown, 1975.
[20]S. M. Pizer et al., "Adaptive histogram equalization and its variations," Computer vision, graphics, and image processing, vol. 39, no. 3, pp. 355-368, 1987.
[21]K. Zuiderveld, "Contrast limited adaptive histogram equalization," in Graphics gems IV, 1994, pp. 474-485.
[22]K. Z. M. Azmi, A. S. A. Ghani, Z. M. Yusof, and Z. Ibrahim, "Natural-based underwater image color enhancement through fusion of swarm-intelligence algorithm," Applied Soft Computing, vol. 85, p. 105810, 2019.
[23]P. Drews, E. Nascimento, F. Moraes, S. Botelho, and M. Campos, "Transmission estimation in underwater single images," in Proceedings of the IEEE international conference on computer vision workshops, 2013, pp. 825-830.
[24]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, 2010.
[25]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, 2015.
[26]J. Tong, T. Zhang, and L. Zhang, "Towards underwater image restoration: A physical-accurate pipeline and a large scale full-reference benchmark," in 2022 IEEE International Conference on Multimedia and Expo (ICME), 2022: IEEE, pp. 01-06.
[27]Y. Wang, J. Zhang, Y. Cao, and Z. Wang, "A deep CNN method for underwater image enhancement," in 2017 IEEE international conference on image processing (ICIP), 2017: IEEE, pp. 1382-1386.
[28]C. Li et al., "An underwater image enhancement benchmark dataset and beyond," IEEE transactions on image processing, vol. 29, pp. 4376-4389, 2019.
[29]M. J. Islam, Y. Xia, and J. Sattar, "Fast underwater image enhancement for improved visual perception," IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 3227-3234, 2020.
[30]C. Li, S. Anwar, J. Hou, R. Cong, C. Guo, and W. Ren, "Underwater image enhancement via medium transmission-guided multi-color space embedding," IEEE Transactions on Image Processing, vol. 30, pp. 4985-5000, 2021.
[31]L. Peng, C. Zhu, and L. Bian, "U-shape transformer for underwater image enhancement," IEEE Transactions on Image Processing, vol. 32, pp. 3066-3079, 2023.
[32]Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, "Image quality assessment: from error visibility to structural similarity," IEEE transactions on image processing, vol. 13, no. 4, pp. 600-612, 2004.
[33]C. W. Park and I. K. Eom, "Underwater image enhancement using adaptive standardization and normalization networks," Engineering Applications of Artificial Intelligence, vol. 127, p. 107445, 2024.
[34]C. Ancuti, C. O. Ancuti, T. Haber, and P. Bekaert, "Enhancing underwater images and videos by fusion," in 2012 IEEE conference on computer vision and pattern recognition, 2012: IEEE, pp. 81-88.
[35]X. Fu, P. Zhuang, Y. Huang, Y. Liao, X.-P. Zhang, and X. Ding, "A retinex-based enhancing approach for single underwater image," in 2014 IEEE international conference on image processing (ICIP), 2014: Ieee, pp. 4572-4576.
[36]X. Fu, Z. Fan, M. Ling, Y. Huang, and X. Ding, "Two-step approach for single underwater image enhancement," in 2017 international symposium on intelligent signal processing and communication systems (ISPACS), 2017: Ieee, pp. 789-794.
[37]C. Li, J. Guo, C. Guo, R. Cong, and J. Gong, "A hybrid method for underwater image correction," Pattern Recognition Letters, vol. 94, pp. 62-67, 2017.
[38]Y.-T. Peng, K. Cao, and P. C. Cosman, "Generalization of the dark channel prior for single image restoration," IEEE Transactions on Image Processing, vol. 27, no. 6, pp. 2856-2868, 2018.
[39]C.-Y. Li, J.-C. Guo, R.-M. Cong, Y.-W. Pang, and B. Wang, "Underwater image enhancement by dehazing with minimum information loss and histogram distribution prior," IEEE Transactions on Image Processing, vol. 25, no. 12, pp. 5664-5677, 2016.
[40]Y.-T. Peng and P. C. Cosman, "Underwater image restoration based on image blurriness and light absorption," IEEE transactions on image processing, vol. 26, no. 4, pp. 1579-1594, 2017.
[41]W. Ren, S. Liu, H. Zhang, J. Pan, X. Cao, and M.-H. Yang, "Single image dehazing via multi-scale convolutional neural networks," in Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, 2016: Springer, pp. 154-169.
[42]K. Panetta, C. Gao, and S. Agaian, "Human-visual-system-inspired underwater image quality measures," IEEE Journal of Oceanic Engineering, vol. 41, no. 3, pp. 541-551, 2015.
[43]M. Yang and A. Sowmya, "An underwater color image quality evaluation metric," IEEE Transactions on Image Processing, vol. 24, no. 12, pp. 6062-6071, 2015.