Research On Batch Modeling Technology Of Tilted Image Data Based On CGA Rules

With the rapid development of geographic information related technologies and applications,the demand for three-dimensional models of urban buildings is growing.It is extremely important how to quickly and batch-complete three-dimensional modeling of buildings,The parametric modeling technology based on semantic rules is very suitable for large-scale batch modeling.In addition,the acquisition of geospatial data sources is constantly updated and efficient.Tilting photogrammetry as a high technology in the field of international surveying and mapping has broken through the limitations of traditional aerial surveying techniques,such as the inability to achieve efficient use of geographic information data,long operating cycle,and other shortcomings.It uses multi-angle cameras for multi-view shooting to obtain full-fledged elements.Detailed geographical information,in order to quickly obtain real and accurate 3D geographic information.Based on the above two aspects,this paper attempts to cross over multiple fields,and proposes a batch modeling technique for oblique image data based on semantic rules.First of all,in the selection of geospatial information data source,this paper abandons the traditional method of acquiring basic building data,and tries to extract buildings automatically from oblique image data.Second,processing oblique image data,and proposes a method for extracting buildings based on DSM derived contours and further extracting buildings from oblique image data.Finally,using batch modeling techniques based on CGA rules,For the geometric data of buildings extracted from the above process,more efficient CGA rules are written to complete real-world,high-precision building batch modelling,thus solving the parametric modeling process.The problem of unreal and accurate building geometry information.At the same time,since most of the tilt models automatically generated by the tilting post-processing platform are triangular mesh models constructed based on point cloud data,it is actually a 2.5-dimensional digital surface,so it also provides a way to realize the singulation of the tilt model.Solutions.Thesis mainly achieved the following results:First,abandon the acquisition methods of traditional geospatial data sources.Starting from the development of cutting-edge photogrammetry technology at the forefront of the development,processing the oblique image data generated by it and attempting to automatically extract buildings from it.Provides a new application exit for tilted image data.Second,innovatively propose a building extraction method based on oblique image data.That is,by processing oblique image data to generate three-band DSM and single-band DOM image,and then based on DSM derived contour algorithm,adjusting different thresholds to optimize,so as to quickly and efficiently obtain the complete topology of the building outline vector data and true and accurate Height information.Third,Write more efficient CGA rules for automated batch modeling based on real geometry data of buildings.Design the CGA rule modeling process.And select the buildings extracted from the experimental area based on the oblique image data as an example to realize semantic rule-driven modeling.On the one hand,it solves the highly inaccurate problem of rule modeling in normal situations.On the other hand,compared to traditional modeling methods,it requires the tedious work of modeling each building separately.CGA rules enable efficient and time-saving modeling.