The Research On Underwater Imaging, Underwater Image Enhancement And Relevant Applications

With the continuous development of the international situation, Ocean has gradually become the new strategic center of the world. The research on oceanography is important for ocean resources exploration and development, marine environmental monitoring and Marine military applications. Against this background, this thesis focuses on the research of underwater imaging, underwater image enhancement, marine bio luminescence, submarine’s light wake and underwater target detection.Underwater visibility is very important in underwater vision research and image processing. However, most underwater vision systems cannot guarantee satisfying performance under complicated water conditions. This is because underwater images are usually degraded by the light-water interactions of absorption and scattering. Background light caused by scattering of ambient light is the main reason of underwater image degradation. In this thesis, the formation of underwater images is analyzed by setting up physical models under two different lighting conditions: natural illumination and artificial lighting. The global background light under natural illumination is proportional to the scattering coefficient and inversely proportional to the attenuation coefficient. The background light under the two lighting conditions both can be expressed in simple exponential falloff expressions of the global background light. The simple expression greatly reduces the computational complexity of simulations. The intensity of background light mainly depends on the inherent optical properties, camera-scene distance, camera-source distance and camera’s imaging angle. The relationship between the global background light and the inherent optical properties can be used to estimate the attenuation coefficients, scattering coefficients and scene depth information. The result of the simulation can be very useful for the design and improvement of underwater imaging systems. Simulation results of underwater images show that the effects of direct attenuation, forward scattering and backscattering on the image degradation mainly depend on the inherent optical properties and camera-object distance.In this thesis, an underwater image enhancement method based on optical attenuation coefficients of different light wavelengths is proposed. Light losses of different wavelengths during the light-water interaction are compensated using different transmission maps of three color channels, which are estimated based on Dark Channel Prior and the relationship between global background light and inherent optical properties. Finally, a color-corrected and haze-free image can be recovered using an underwater image formation model. The main contribution of the method is to estimate the attenuation coefficient ratios between different colors of light using the global background light from a single image. Quantitative and qualitative comparison experiments show that the proposed method improves color rendition and removes haze simultaneously with good performance.Based on the research of bioluminescent marine creatures’luminescence characteristics and CFD. simulations of submarine propeller’s stress field, the formation mechanism and characteristic of ship’s light wake are studied. The shape and size of light wake can be approximately estimated using luminescent response function of fluid shear stress and submarine propeller’s stress distribution. Simulation results show that the ship’s light wake can be hundreds of meters long; it is very convenient to detect and track submarine in the air or underwater. Finally, the system configuration of light wake’s detection and tracking is presented.