Research On Underwater Image Restoration And Mosaic Technology For AUV Application

With the rapid development of ocean research,the application of underwater optical vision systems has been expanded vastly.However,due to the absorption and scattering of light by the water environment,underwater images often have quality degradation problems such as blurring and color distortion,and the attenuation of light also seriously affects the field range of underwater imaging.To a large extent,these problems restrict the usefulness of optical vision system in underwater operations.This paper mainly focuses on the above issues,and completed works are as follows:(1)The transmission property of light in water and the imaging model are analyzed.On this basis,an underwater image clearness algorithm based on dark channel prior is proposed.According to the characteristics of underwater imaging,the algorithm improves the estimation of ambient light and light transmittance.At last,validity of this proposed method is verified by experiments.(2)Considering the small field range of underwater imaging,this paper applies image mosaic technology to underwater environment.And panoramic image of underwater scene can be obtained from video image sequence.In order to improve the speed of mosaic,an image overlap region estimation method based on perceptual hash theory is proposed.This method can effectively narrow the target range of feature detection and matching.Also,in order to improve the accuracy of image matching,texture features are combined with SURF to enrich the description of features.(3)For improving the convenience of vision system on AUV,an underwater image acquisition and transmission system is designed,and correspondingly,a software is designed on PC,which realizes the underwater video monitoring function,along with the processing and mosaic function of video images...