A group of specialists from Peter the Great St. Petersburg Polytechnic University, East China Normal University, and the Indian Institute of Technology Roorkee are working together to develop software that will create detailed 3D models of urban infrastructure objects using a combination of laser scanning and photography. The project has become the winner of the international scientific consortiums competition of the BRICS countries in the field of science, technology, and innovation (BRICS STI Framework Programme).
The specialists are going to develop a software to analyze 3D images of infrastructure objects. The data is collected using the laser scanner survey technologies and photography. The software will be useful at all stages of the urban infrastructure lifecycle, including the detection of defects, building information modeling (BIM) and carrying out renovation work.
The first ideas on hybrid processing were laid in 1986 – the method of “absolute orientation” in two coordinate systems simultaneously. This method uses a quaternion (4×4 matrix) to transform a three-dimensional body between two systems in one operation, however, low performance and low availability of computing devices made this method difficult to apply. The term “hybrid processing” assumed the use of joint processing of two sets of 3-dimensional data by the method of “Iterative nearest-point algorithm” to accelerate the finding of local minima using the “annealing simulation algorithm”. The final term “hybrid processing” assumes simultaneous usage of data from a laser scanner and camera to reduce noise and improve the quality of surface reconstruction.
“In cooperation with our foreign colleagues, we analyze the point clouds, obtained by laser scanner survey technologies and develop the software that enables to recognize the particular features of the materials, the texture of the objects and its changes. Our developments can be used for monitoring and repair of the road infrastructure, as well as on the preservation of objects of cultural and historical heritage”, said Vladimir Badenko, professor of Institute of Civil Engineering SPbPU.
The international team of scientists develops the tools for processing the raw data of laser scanning survey, at the first stage transforming the information about the object into the point clouds of suitable density. For the point clouds’ visualization, researchers create the image projections, combining it with the photography data. It significantly improves the quality of laser scanning point clouds segmentation. Using such projections for object modeling enables to display all three-dimensional data on the plane (on the monitor) and, accordingly, to improve the quality and processing speed. In addition, it helps to adjust the required display scale of the object and to obtain the most accurate image (without shadows).
It should be noted, that a significant amount of data is processed using the facilities of the Supercomputer Center “Polytechnic” at Peter the Great St. Petersburg Polytechnic University.