Abstract: Developed at KSU named after K.E. Tsiolkovski's augmented reality system allows for interactive interaction with 3-dimensional models that visualize significant objects of the urban environment of the city of Kaluga. The implementation of this interaction consists in segmenting the image of the layout, recognizing fragments in the segmented image, replacing image fragments with 3D models of objects, and subsequent interactive interaction with them – rotation, scaling, and translation. An additional non-interactive layer allows you to display information related to the object. The augmented reality system is implemented as an application for the iPad and is currently used for a deeper immersion in the urban environment in order to familiarize university guests with the municipal, cultural and historical sites of the city of Kaluga.
Index terms: augmented reality, convolutional neural network.