| Underwater robots,as the only carrier capable of working in deep-sea environments,play an irreplaceable role in developing marine resources and ensuring coastal defense safety.Among them,the Autonomous Underwater Vehicle(AUV),which has the characteristics of flexibility and wide working range,has received widespread attention.In the dark and complex deep-sea underwater environment,AUVs based on visual image operations need to collect images of underwater scenes with the assistance of artificial light sources.Unlike natural lighting environments,artificial light sources can experience severe scattering in aqueous media,resulting in severe degradation of images such as reduced contrast,color deviation,and fog cover.At the same time,artificial light has limited intensity and illumination distance,making most of the background area a dark background,Directly affects the effectiveness of image based AUV operations.Therefore,in order to provide AUV with clearer underwater image information,this paper studies a restoration method suitable for underwater images in dark backgrounds under artificial lighting.It has important research significance and practical value for improving the perception ability of underwater equipment to the environment.This article takes AUV based visual image operations in the deep sea as the research background to study underwater image restoration methods for dark backgrounds under artificial lighting.Conduct theoretical and experimental research on the methods for solving depth of field,background light,and underwater attenuation coefficient in the process of underwater target image restoration.Firstly,this article investigates the estimation method of depth of field in underwater image restoration under artificial illumination.In response to the typical scene depth estimation methods used in the past,when applied to underwater image restoration of dark water background under artificial lighting,there is a significant difference in the depth of field obtained from the same distance target,which leads to distortion in the brightness(grayscale)of the restored image and affects the quality of image restoration.This paper studies and improves a depth of field estimation method based on image blur.The effectiveness and applicability of the proposed method were verified through comparative experiments on the depth of field estimation results of multiple scenes.Secondly,this article investigates the estimation method of background light in underwater image restoration under artificial illumination.When applying to underwater image restoration of dark water background under artificial lighting,the obtained background light is too large,resulting in significant distortion of the overall image brightness after restoration,which affects the quality of image restoration.This article studies and improves a background light estimation method based on depth of field segmentation.The comparative experiments of multiple scene background light estimation results have verified that the proposed method has better environmental adaptability compared to existing typical methods.Then,this article studies the estimation method of transmittance in underwater image restoration under artificial lighting.Due to the unknown turbid water environment under artificial illumination of AUV,the classical transmittance estimation method based on the assumption of "attenuation coefficient can be obtained from experience" in the past has a significant difference between the R,G,and B transmittance obtained in the underwater scene in this paper and the actual attenuation,resulting in significant color distortion of the restored image and affecting the quality of image restoration.This article proposes an underwater image transmittance estimation method based on dark water background under artificial lighting.Through comparative experiments on the transmission estimation results of multiple scenes,it has been proven that the method proposed in this paper is more suitable for the transmission estimation of underwater images in dark water backgrounds under artificial lighting.Finally,an experimental study was conducted on underwater image restoration of dark water bodies under artificial lighting.In response to the artificial illumination of dark water background underwater images collected in this experiment,the depth of field estimation method,background light estimation method,and transmittance estimation method proposed in this article were used to study the restoration effect of underwater images collected from different distances and target objects.Through a comparative experiment with existing classic underwater image restoration methods,Verified the effectiveness and applicability of the underwater image restoration method based on artificial lighting proposed in this article. |
