Nonuniformly illuminated image enhancement based on range-gated underwater imaging system

Range-gated underwater laser imaging technique can eliminate most of the backscattering and absorption noise effectively. It has a range of from 4 to 6 times that of a conventional camera with floodlights in the strongly scattering waters, which becomes a useful technique in oceanic research, deep-sea exploration, underwater remote control and robotic works. While because of the laser pulse stretching, the image obtained through range gated underwater imaging system has obvious nonuniformly illuminated character, such as brighter center and darker edge. Low contrast and grayish white of the image also bring great difficulty for processing. In order to adjust the lightness of the nonuniformly illuminated image of range-gated underwater imaging system, the water degradation is assumed as illumination variation and retinal-cortex theory based on color constancy is introduced. Frame integral algorithm has to be applied first to eliminate system noise for the reason that we adopt the system based on intensified gate imaging technology. And gray stretch ensures that we can attain appropriate output. In retinal-cortex models, McCann model and McCann-Frankle model have obvious effect. So we choose the two models for comparison and improve the second one considering the exponential characteristics of eyes for illumination. In order to evaluate the methods objectively, strength uniformity of signals is applied. The experimental results demonstrate that the approaches we adopted are all effective and can enhance the image contrast. And the improved McCann-Frankle model gets more satisfying visual effect.