Augmented reality as a tool for virtual geographic environments

Abstract

A virtual geographic environment, as a viewing and interaction tool, facilitates the interpretation of special data, contrary to the experience on flat figures. We present a method to fuse and view digital surfaces and ortophotos over topographic planes by using augmented reality techniques adapted to be used with georeferenced data and facilitating human-machine interaction in geographical works. We developed a Matlab tool to integrate surfaces and planes by tracking feature points in the color space HSV with neural networks. We designed a test protocol and assessed the tool by using a dataset composed by point clouds and ortophotos from a section of the municipality Versalles, department Valle del Cauca in Colombia. The evaluation was qualitative for user experience by assessing different virtual elements like: appearance, required illumination, etc. The evaluation was quantitative to test the viewing precision by the integration method. The tests show a user satisfaction of 91.48% for static images and a mean error of 6° for azimuth and 3.5° for elevation. The evaluation shows that the developed method facilitates the interpretation of georeferenced ortophotos.