Research On Storage Management And Index Of The Massive Texture Image

T-Li DAR is a high-tech instrument of surveying and mapping with a combination of advanced technologies. It is well known as the 3D laser scanning technology in surveying and mapping field for the features of high efficiency, precision and non-contact measurement. As one of the main methods of obtaining refined 3D spatial information, T-Li DAR technology is widely applied in the protection of ancient architecture, 3D modeling, the large complex buildings’ construction quality, deformation monitoring and disaster assessment survey etc.With the rapid promotion in Li DAR manufacturing technique, the updated hardware and software bears a better capacity for the increasing data acquisition. In order to obtain a more accurate model, project operators attempt to acquire more dense original data. Massive original data, including point clouds and images, etc., gets larger and more complicated after the subsequent data processing and continuously data update, thus the data storage and management technology must be improved to better serve the development of the 3D laser scanning technology. During the process of the project, the quantity of Hou Mu Wu Ding’s body and partial photos hits 2,000 and takes 500 M storage space. Extra storage space is as well required to meet the need of data copy and processing when deleting and adding images. Apparently, the traditional image management technology, based on computer storage, can no longer meet the requirements in data storage and management against the increasing massive data. The author, based on the current image management technology, studies the storage, management and visualization of the image files in Postgre SQL database management system combined with the file management technology of the object-relational database system.Firstly, the object-oriented Standard Template Library（STL） and C/C++ Language（file operation supported）, by analyzing the requirements for image data management, are applied to develop the experimental system. The system developed can transfer the image files to block data shortly after the speed extraction. Meanwhile, data blocks are stored as byte in the field of BLOB type and the properties information parsed（time, exposure and etc.） are stored in database. The author as well introduces the features of the object-relational database and those of the image files in multiple formats. The author verifies Lib-PQ as the access portal by analyzing the features of the access portal and data management mechanism of the database. Thirdly, the author confirms the appropriate tree as the image index and stores the tree in a sheet of the database by analyzing different index trees to fit the requirements of processing 3D modeling image for buildings. Finally, the real-time updates of the sheets and index trees can be managed by using the built-in triggers in Postgre SQL database.In this article, the programming is completed under the CLR environment by using C++ Language based on the development platform of Visual Studio 2013. After the programming, data could be managed in Postgre SQL database. The software L&P₃D V1.0 which is developed by our team has many modules. The management module of image data is part of the software. It is designed with the functions of open image, data storage, index build, location matching, search and display function and etc. The humanized interaction design of the software further optimizes the generation of the node tree, data update efficiency and other functions. During the software development, a new feature that the tree index structure of images displaying on the user interface improves the readability. Speed searching can be implemented by clicking or selecting the image within the scope of the image index tree. The author tests the feasibility of the management module of image data by processing the data acquired from the team’s projects. The result shows that the module can greatly improve the data processing efficiency, which simultaneously verifies the practical significance of the research content in this thesis.