Method Of Underwater Image Restoration Based On Active Illumination

With the increasing emphasis on underwater resources,the demand for underwater image recognition,target detection and others becomes higher and higher.However,due to the serious degradation of underwater images,the definition of underwater images is low,and the contrast is reduced.The target information is submerged in noise.The acquired underwater images are often not directly applicable to the fields of underwater image recognition and target detection.The underwater image processing method has the function of image restoration and enhancement,which can improve the quality of underwater images.As the prerequisites for image recognition and target detection methods,underwater image processing methods are getting more and more attention.Due to the particularity of the underwater environment,the existing underwater optical image processing methods have great limitations,and cannot obtain good image processing results.Therefore,in order to improve underwater image processing effects,this paper intends to study underwater image processing methods.The specific content is as follows:1.Based on the classical underwater imaging physical model,the light propagation process is analyzed in detail,the backscattering and absorption are considered,and the infinite ambient light estimation is avoided.The improved underwater imaging physical model is proposed,which provides the basis for underwater image restoration later.The underwater imaging experiment platform was designed and built to obtain underwater images under different turbidity and different targets.2.Multi-parameter optimal reconstruction underwater polarization image restoration method is proposed.The method is based on the analysis of the influence of different parameters on underwater imaging.By changing the parameter values and adjusting the image restoration effect,the underwater image restoration is realized.Experiments show that the proposed method can effectively improve the underwater image quality,especially in the condition of high turbidity.Under certain conditions,the restored image contrast is improved by 40% compared with the contrast algorithm.This method has a good restoration effect on the center of the image,but it lacks the details of the edge of the image.Because it needs to go through several iterations,and each iteration needs to recalculate the image processing results on all pixels,the calculation speed of this method is slow.3.An underwater image restoration method based on absorption coefficient and backscattering coefficient is proposed.This method selects multiple target points,and by analyzing the relationship between the target points,achieves the optimal solution of absorption coefficient and backscattering coefficient.According to the optimal solution,the backscattered light is removed and the lost signal light is recovered to restore the underwater image.Experiments show that the proposed method can effectively improve the underwater image quality,especially in the condition of high turbidity.Under certain conditions,the restored image contrast is improved by 25%compared with the contrast algorithm.The processing speed of this method is fast.Due to the more comprehensive consideration of image information,the over-processing situation is not easy to occur,the overall grasp of the image is better,the recovery of image edge details is more perfect,and the overall quality of the image is higher.