基於簡化水下成像模型與高信賴度景深圖之水下影像修復研究

Underwater Image Restoration Based on a Simplified Image Formation Model with a Reliable Scene Depth Map

官品宜(Pin-Yi Kuan)

指導教授 ： 張恆華

國立臺灣大學/工學院/工程科學及海洋工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202100965

摘要

水下影像經常出現顏色失真和低能見度的問題。在真實場景中，光在通過水的時候會衰減，而水下的衰減量根據其波長的不同而不同。本篇論文提出一新的水下影像修復技術，其結合了簡化的傑夫-葛萊蒙利(Jaffe-McGlamery) 水下光學模型和簡化的朗伯(Lambertian)模型。首先推導出影像與發光強度之間的關係，並改進發光強度估算方法以補償光源衰減。接著，提出強化的景深估量方法和改良的彩色邊緣檢測法將前景物體與背景分離，並使用前景特徵與背景之間的相關性來估計背景光。最後，使用線法估計透射圖，並且使用估計的整體顏色增益來校正顏色失真。本論文搜集一系列不同場景和色偏的水下影像，並使用它們來評估所提出之方法。實驗的數據指出，在各種水下場景的修復中，本論文提出的演算法相較許多現存方法擁有更穩定的色偏修正和去散射能力，顯示出本方法在許多水下影像修復應用之潛力。

關鍵字

水下影像 ； 影像修復 ； 發光強度 ； 光源補償 ； 景深 ； 彩色邊緣檢測

並列摘要

Underwater images often suffer from color distortion and low visibility. In reality, light attenuates when passing through water and the amount of attenuation varies according to its wavelength. This thesis combined the Jaffe-McGlamery and simple Lambertian models and derived the relation between underwater images and the corresponding illuminant intensity. A modified illuminant intensity estimation method was introduced to compensate for the light source attenuation. Then, a modified scene depth method and a modified color edge detection scheme were used to separate the foreground objects from the background, form which the correlation between the foreground attributes and the background was used to estimate the background light. Finally, the haze lines method was used to estimate the transmission map and the estimated ensemble color gain was used to correct the color distortion. A wide variety of underwater images with different scenarios were collected and adopted to evaluate the proposed algorithm. Comparing with many state-of-the-art methods, the proposed method had more stable color distortion correction and de-scattering capabilities, which indicated its potential in many underwater image restoration applications.

並列關鍵字

underwater image ； image restoration ； illuminant intensity ； light source compensation ； scene depth ； color edge detection

參考文獻

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.

D. Akkaynak, T. Treibitz, T. Shlesinger, Y. Loya, R. Tamir, and D. Iluz, "What is the space of attenuation coefficients in underwater computer vision?," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 4931-4940.

N. Jerlov, "Irradiance optical classification," in Optical Oceanography: Elsevier, 1968, pp. 118-120.

I. Kashif, A. Rosalina, O. Azam, and Z. Abdullah, "Underwater image enhancement using an integrated colour model," IAENG International Journal of Computer Science, vol. 34, no. 2, 2007.

A. S. A. Ghani and N. A. M. Isa, "Underwater image quality enhancement through integrated color model with Rayleigh distribution," Applied soft computing, vol. 27, pp. 219-230, 2015.