3D Spatial Arrangement Method of 360° Image with Depth Applied Considering Player Movement

Abstract

The game environment similar to the real environment increases the user's sense of immersion. Especially in the real world, buildings are an important factor in giving users a sense of realism. Real environmental data such as satellite image, aerial image, and 360° image have already been digitized and are being serviced like Google Street View or Kakao Road View, but there is no special reprocessing. Constructing the world from 3D models rather than images is already a work in progress. However, the work done by Google and others mainly uses aerial images as textures for their models, which are highly distorted and have low resolution. In this study, an independent building mesh was created in 3D space by extracting street images and metadata from the existing 360° street image service to compose a map where users move and interact. Street views are placed in a 3D space using TM coordinates so that appropriate street view is applied according to the user's movement. We analyzed the depth data to apply depth to a spherical screen projected into street view, separated the buildings from the background in street view, and projected each building onto a separate vertical plane. Each building was also rotated to maintain its initial orientation following the user, giving the impression of continuity of the building plane in transitions between street views. This study aims to increase the value of existing digital data by reprocessing it so that it can be used in various fields.