J. Judvaitis, I. Homjakovs, R. Cacurs, K. Nesenbergs, A. Hermanis, K. Sudars, Object transparent vision combining multiple images from different views, Automatic Control and Computer Sciences, Volume 49, Issue 5, pp. 313-320, 2015. doi: DOIhttps://doi.org/10.3103/S0146411615050053

All authors:

Janis Judvaitis, Igors Homjakovs, Ricards Cacurs, Krisjanis Nesenbergs, Atis Hermanis, Kaspars Sudars

Abstract:

Augmented reality (AR) has been studied extensively and used widely in various fields. In AR a view of reality is modified or supplemented by computer-gene-rated sensory input such as sound, video, etc. As a result, it can enhance one’s current perception of reality. In this paper we analyzed methods to make objects transparent by fusing images taken from two different cameras. Induced transparency could be particularly interesting for traffic applications. When two cars are driving one after another and the car in front can block the view for the second car, however, if both cars are equipped with cameras capturing the front view, their images can be merged and the front car can be made transparent allowing driver to have clear picture about obstacles and other dangers on the road. The goal of this paper is to analyze various solutions that can provide such transparency effect and also to create a solution that can work in real-time.