Three-Dimenson Ship Reconstruction Based On Binocular Stereo Vision

With the rapid development of shipbuilding industry,intelligent ship has become one of the main focuses of the current development of shipbuilding industry,and the intellectualization and automation of shipbuilding industry can not be separated from the support of computer vision.Binocular stereo vision is an important branch in machine vision.It is based on parallax principle and uses imaging equipment to obtain two images of the object under test from different positions.By calculating the position deviation between the corresponding points of the image,it can obtain the three-dimensional geometric information of the object.Binocular stereo vision has the characteristics of convenience,low cost,strong real-time performance and more appropriate accuracy.It has been widely used in robotic vision,aerial mapping,reverse engineering,military applications,medical imaging and industrial testing.At the same time,with the development and maturity of SLAM,the registration of point clouds no longer requires auxiliary means such as markers,which creates the possibility for real-time three-dimensional reconstruction of large-scale objects such as ships.On the basis of binocular stereo vision and visual SLAM,a fast threedimensional reconstruction system based on binocular stereo vision is designed to meet the accuracy requirements of ship's overall performance calculation.This system not only makes it possible to obtain ship's three-dimensional model quickly,but also provides a more accurate and rapid method for ship's construction precision control,ship damage assessment and navigation of marine robots route.In this paper,binocular stereo matching and SLAM technology will be studied in depth.Based on the accelerated algorithm of SUDA,the improved algorithm of ORB-SLAM2 dense reconstruction and the framework of QT,a practical binocular ship 3D reconstruction solution system will be designed and implemented.