Examination And Realization On Three-dimensional Reconstruction Algorithm Of Supercavitation

In recent years, three-dimensional reconstruction has become a hot research field of computer vision for its wide range of applications. Along with the deep-going study of theoretical knowledge and improvements of technique, it effectively promotes the development of technology in the field of optical imaging, digital image processing, and software programming and designing. In this paper, it estimates the study on the reconstruction algorithm and the key parameters in this algorithm by the two-dimensional grayscale image of supercavitation under water with the background of a project for a research institution.At the beginning of this thesis, the background information, significance and basic principles of three-dimensional reconstruction are introduced, including the research status at home and abroad at present. It selects the method of shape from shading(SFS) to do the three-dimensional reconstruction of the supercavitation's image because of the special nature of this topic. After analyzing and comparing the pros and cons of the common illumination models and commonly used SFS algorithms of reconstruction, it confirms the scheme of three-dimensional reconstruction and flow path of algorithm for the grayscale image of supercavitation. Then it analyzes the influence on the result of reconstruction when selecting the position of light source and the value of slant and tilt of point on the surface of object in optical imaging. On this basis, a mathematical solution for supercavitation on SFS algorithm is deduced and the methods of the pre-processing are discussed for the grayscale image of supercavitation and the model of natural supercavitation which support the analysis of precision for the results of reconstruction. At last, it completes the SFS algorithm for three-dimensional reconstruction on the grayscale gradient image of supercavitation in the environment of MATLAB. In order to verify the accuracy of algorithm, firstly it selects a grayscale image of the hemisphere under ideal lighting conditions to do three-dimensional reconstruction, and the result is ideal. Subsequently, in the content of research, it builds a test environment under laboratory conditions, takes photos of supercavitation which is simulated by manual work under water to do the three-dimensional reconstruction and analysis of results, and points out the aspects that need to be improved in the follow-up work of this project. At the same time, the software system in the environment of MATLAB completes a solver and storage for the three-dimensional parametric of grayscale image, which provides related data for the dynamic visual simulation of supercavitation in future.