Evaluations on 3D Personal Navigation based on Geocoded Images in Smartphones

Abstract

3D personal navigation is becoming a standard feature in smartphone platform, which develops in a fast speed nowadays. However, the hardware restrictions of smartphone may degrade the 3D rendering performance, and such real-time operation is not an energy-efficient procedure on smartphone, because heavy computation consumes a lot of power, which is crucial for a smartphone equipped with limited capacity battery. This paper presents a novel solution utilizing geocoded images instead of 3D models to mitigate these technical restrictions on the smartphone. To demonstrate the performance and the improvement of the proposed solution, evaluations are carried out in term of positioning accuracy, resource consumption, efficiency, visualization, and labour costs. The results show that the proposed solution has overwhelming advantages in all these comparisons. This solution also has the capability of achieving a higher frame rate and has a better visualization performance as well. In addition, the proposed solution provides an optional way to decrease the labour costs and hardware investment to build up a similar but quick application by utilizing photos instead of complex 3D model construction for a small-scale area personal navigation application.