The Research Of Mixed Light Field Underwater Optical Two-dimensional And Three-dimensional Detection

Marine science and technology has become the focus of world with the economicand technological development. For the purpose of underwater exploration andconstruction, wide range and accurate underwater environmental data are needed. Inorder to obtain the data, a series of ocean technologies were under researched. And theunderwater imaging method is researched in this article.This paper presents a new approach for underwater target detection byinhomogeneous illumination, which intensity distribution relates to the attenuationpattern of light transmission with an inverse form. The proposed approach was testedand verified by the underwater imaging experiments, showing that it is effective todecrease the backscattering within a wide field of view (FOV) to increase thedetection distance, especially through turbid water. By the proposed method,underwater target detection with wide FOV, low backscattering noise (especiallythrough turbid water) and non-blind area can be realized.Laser line scan (LLS) is one of the underwater optical imaging methods, whichcan reduce the detrimental effects of backscatter and forward scatter. By projectinglaser stripe onto the target, the profile of target can be inferred from the displacementof the laser scan lines.Inhomogeneous light field image detection system was used in the poolexperiments. The underwater photometer was use to measure the energy distributionof the axis of the receiver to the target (LOS). While get the pictures of the targetunder different distances. It can concluded that inhomogeneous light field imagedetection system in accordance with the design theory, and can improve the detectionrange effectively. Experiments had been carried out in a sink, and the scanning system was set up in alinear guide. A532nm semiconductor laser was used as light source, and a cylindricallens was placed ahead of it to generate a laser line. The2D data were acquired by adigital camera. Some image processing methods were used for denoising and centerextraction. Then the coordinates of2D data were transformed into3D worldcoordinates. And the3D reconstruction result is shown at millimeter scales.