Currently, most of cyber city systems utilize low level of detail building models. This research develops systematic methods to increase the level of detail of OGC (OpenGeospatial Consortium) LOD2 building models using close-range images. The developed algorithms use image features and characteristics of building structures in real world with semantic analysis and digital image processing to recognize the structure objects like windows, doors, balconies, etc. Secondly, space intersection is employed to determine the dimensions and positions of the structure objects. The identified objects are then added into existing building models to increase their level of detail. The developed methods can be used to generate 3D building models with higher level of detail more effectively and efficiently.