3D Reconstruction Of Deep Sea Area Based On Line Structured Light

The deep sea is the largest unknown area on the earth,which contains all kinds of strategic materials and energy for the development of human society.The realization of high-precision 3D reconstruction of submarine topography can provide reliable basis for underwater operations,such as target recognition,geological exploration,mineral resource estimation,environmental monitoring and submarine path planning.This paper base on the principle of 3D reconstruction with line structured light method,focus on the hot research areas(cold spring area,hydrothermal area and mineral area),design a vision sensor composed of line laser and camera,which integrate the navigation-positioning data and attitude information of Jiaolong manned submersible to build the 3D reconstruction system of deep sea.Specific research contents include:(1)In this paper,according to the characteristics of the underwater image,the basic image processing algorithm is discussed.For the purpose of image preprocessing,this paper select the Gaussian filter to reduce the noise of the image,ues the corrosion expansion algorithm to reduce the impact of the suspended solids,and chose the gamma correction to enhance the image characteristics.(2)The accuracy of extracting structure light stripe centerline determines the precision of 3D reconstruction.Through the improved Steger algorithm,the fast extraction of subpixel centerline is realized in this paper,and the program test shows that the accuracy of extraction is more than 98%.(3)Camera calibration is the key to reflect the real target characteristics of the 3D reconstruction system.According to the projective geometry characteristics,this paper use the direct calibration method to calculate the image coordinates,and combine with the displacement of the visual sensor for achieving the transformation of pixel coordinates to the real world coordinates.The accuracy of the direct calibration method is up to 97% by the experiment validation.(4)In order to ensure the continuity and accuracy of the terrain teconstruction in the whole region,the image coordinates obtain by visual sensors are integrated with the position information of the manned submersible,thus unifying them to the fixed world coordinate system.According to the attitude information of the manned submersible,the calibration surface and reconstruction coordinates are corrected to make the acquired pointcloud coordinates follow the trajectory of the submersible and realize theaccurate 3D reconstruction of the seabed area.According to the designed 3D reconstruction system,the underwater simulation experiment platform is built.After the experiment and program verification,the accuracy of the reconstruction is more than 95% and which is analyzed the reason of the error.Considering the high-pressure environment in the deep sea,this paper design the titanium alloy package of the laser and select the appropriate underwater pressure camera to form the visual sensors,which meet the carrying requirements.Finally,the system running process is designed to lay the foundation for the practical applications.