Research On 3D Reconstruction Technology Of Underwater Target Based On Point Cloud Data

ROV is widely used in offshore oil and gas development,marine scientific investigation and other fields.However,because ROV belongs to remote control operation,it is difficult for operators to perceive the relative position and underwater operation situation of ROV in real time,which affects the safety and efficiency of ROV operation.This paper takes ROV underwater operation scene as the main research object,and focuses on the research of 3D reconstruction technology of underwater complex environment.(1)In the optimization processing of underwater image sequence and other data sources,this paper analyzes the color gamut space information in the data source,and adds the hybrid filtering algorithm and image threshold segmentation algorithm,establishes an optimization segmentation method of target object image sequence,and realizes the optimization processing of underwater data source.(2)In the optimization of point cloud data generation,this paper designs a set of optimization system of point cloud data generation based on object image sequence by using the improved scale invariant feature transformation matching algorithm and adding the error matching elimination algorithm,which realizes the optimization of point cloud data generation of underwater data source.(3)In the point cloud data surface reconstruction,this paper uses Poisson surface reconstruction method to compare and analyze the main factors affecting the 3D model,and realizes the 3D reconstruction of point cloud data.The optimized segmentation method of target image sequence established in this paper enhances the feature information and edge information of the target in the image,optimizes the effect of image threshold segmentation,and provides a high-quality data source for the threedimensional reconstruction of underwater target.The feature matching system of target image sequence based on image threshold segmentation designed in this paper verifies its better performance in feature point detection The efficiency and robustness of other methods further improve the quality of point cloud data generated by the system.