不同光照条件下水下成像背景光的建模与研究

由于光在水中传输时的衰减和散射效应, 水下成像系统通常很难达到令人满意的成像效果, 而成像过程中由于光散射产生的背景光则是造成水下图像退化的主要原因. 本文对自然光照和人工光照两种不同光照条件下的水下成像背景光进行了建模和仿真分析. 结果显示: 自然光照下的无穷远处背景光与衰减系数成反比, 与散射系数成正比; 人工光照下的无穷远处背景光则与衰减系数、散射系数和相机-光源距离有关. 两种光照条件下的背景光都可以用无穷远处背景光的指数衰减表达式来表示. 水下图像背景光的强度主要与水体光学参数、相机-景物的距离、相机-光源的距离及相机成像角等因素有关. 本文的研究结论不仅可用于水下成像系统的设计与优化, 也可用来从水下图像背景光中估计水体光学参数、相机-景物距离等图像信息.

Modeling and simulation of the background light in underwater imaging under different illumination conditions

The underwater visibility is very important in underwater vision research and target detection. However, most underwater vision systems cannot guarantee to possess the performance under complex water conditions. This is because underwater images are usually degraded by light-water interactions of absorption and scattering. The ambient light is scattered into the camera’s line of sight by water molecules and suspended particles in the water medium, which adds a layer of haze to the image and reduces the contrast of the image. This part of scattered light is usually called background light, which is the main reason for underwater image degradation. In this paper, the formations of background light in underwater imaging under two different lighting conditions: natural illumination and artificial lighting, are analyzed by setting up physical models. The models developed include the parameters such as camera parameters, light source parameters, inherent optical properties, and camera-source-object geometry. Based on the models, the relationship between the background light and the above parameters is studied. Computer analysis shows that the global background light under two illumination conditions has a close relationship between the inherent optical properties of water medium and camera parameters. 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 illumination conditions both can be described 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 coefficient, scattering coefficient and scene depth information. The result of this paper can be very useful for designing and improving the underwater imaging systems.