Underwater Optical Image Processing: A Comprehensive Review (1702.03600v1)

Abstract: Underwater cameras are widely used to observe the sea floor. They are usually included in autonomous underwater vehicles, unmanned underwater vehicles, and in situ ocean sensor networks. Despite being an important sensor for monitoring underwater scenes, there exist many issues with recent underwater camera sensors. Because of lights transportation characteristics in water and the biological activity at the sea floor, the acquired underwater images often suffer from scatters and large amounts of noise. Over the last five years, many methods have been proposed to overcome traditional underwater imaging problems. This paper aims to review the state-of-the-art techniques in underwater image processing by highlighting the contributions and challenges presented in over 40 papers. We present an overview of various underwater image processing approaches, such as underwater image descattering, underwater image color restoration, and underwater image quality assessments. Finally, we summarize the future trends and challenges in designing and processing underwater imaging sensors.

• The paper reviews challenges in underwater imaging like absorption and scattering, offering an exhaustive evaluation of techniques developed over the last five years from over 40 state-of-the-art studies.
• The review categorizes methods into hardware-based techniques like polarization and range-gated imaging and software-based approaches focusing on de-scattering and color restoration.
• It discusses various image quality assessment indices and highlights the research implications for AUVs/UUVs and future directions including deep learning and IoT integration.

Underwater Optical Image Processing Techniques: A Comprehensive Review

The field of underwater optical image processing is characterized by several unique challenges, necessitating advanced techniques to effectively manage issues arising from light absorption, scattering, and non-uniform lighting in underwater environments. The paper, "Underwater Optical Image Processing: A Comprehensive Review," offers an exhaustive evaluation of developments in underwater image processing over the past five years, drawing insights from over 40 state-of-the-art studies. This review is pivotal for researchers and practitioners working with underwater visual data, particularly those involved with autonomous underwater vehicles (AUVs) and unmanned underwater vehicles (UUVs).

Challenges of Underwater Imaging

Underwater imaging presents specific inherent challenges unlike those encountered in terrestrial environments. Scattering due to medium non-uniformity, wavelength-dependent absorption, and artificial lighting effects—such as vignetting and flickering—create obstacles that demand specialized processing methods. These factors not only degrade image quality but also complicate the tasks related to underwater navigation and observation for AUVs and UUVs. Addressing these issues is critical for enhancing both visual quality and the information extracted from marine environments.

Imaging Techniques and Approaches

The paper categorizes underwater image processing methods into hardware-based and software-based approaches. Each approach is delineated into relevant sub-categories that illustrate their unique principles and applicability:

1. Hardware-Based Methods:
   • Polarization Imaging: This method utilizes polarizers to mitigate backscatter, enhancing image clarity significantly.
   • Range-Gated Imaging: Predominantly employed in laser-imaging systems, these methods selectively filter reflected light from objects, blocking backscatter.
   • Fluorescence Imaging: By leveraging the fluorescent properties of certain organisms and materials, this technique aids in shape and microorganism detection.
   • Stereo Imaging: Enhancements in visibility through real-time algorithms applied in AUVs are achieved by estimating visibility coefficients.
2. Software-Based Methods focus on de-scattering and color restoration:
   • De-scattering: Includes both physical model and non-physical model approaches. Techniques such as Markov Random Field methods for haze estimation, wavelength compensation, and dehazing remain central.
   • Color Restoration: Methods address the challenges of light absorption recovery, utilizing models like the spectral response function to restore color contrast and mitigate vignetting effects.

Underwater Image Quality Assessment

The quality assessment of underwater images ensures that processing methods not only enhance visibility but also maintain structural integrity. The review details various quality indices, both reference-based and non-reference, examining image parameters such as colorfulness, contrast, and sharpness. Structural similarity indices (SSIM) and bespoke metrics like UIQM and UCIQE quantify image quality, while novel indices like Qu evaluate structural and color similarities.

Implications and Future Directions

The implications of this research are far-reaching, particularly with the advent of sophisticated AUVs and UUVs that rely heavily on accurately processed optical data for navigation and exploration. The detailed review of methods provides a roadmap for addressing current limitations and improving imaging capabilities in turbid marine environments.

Looking ahead, the paper outlines potential areas for further research, including the integration of deep learning and cloud-based methodologies to advance imaging techniques. Furthermore, the rising confluence of the Internet of Things with underwater imaging promises enhanced real-time processing and data transfer capabilities.

In conclusion, this comprehensive review serves as a critical resource for researchers and engineers in the field of underwater imaging, offering both a foundation and a springboard for continued innovation in overcoming the intricate challenges presented by underwater environments.