Early work on the development of GPSView, an augmented global positioning system (GPS) that returns an optimal scenic driving route, is presented. This is in contrast to commonly used systems for automobiles that provide the shortest route with the least traveling distance. The work in this paper aims to produce both optimal traveling distance and a pleasurable sightseeing experience.
For this investigation, the authors built a database of GPS-tagged scenic landscape photos that are freely available online. By analyzing the geospatial distribution of these images in a mapping system, the authors locate “hotspots” or points of interest (POIs). The best drive-by landscapes are chosen and the route is optimized for both traveling distance and best scenic drive.
However, Google Maps (http://maps.google.com/), Google Earth (http://www.google.com/earth/), and apps such as the free “Navigon” (http://www.navigon.com) on my Samsung Galaxy device already provide similar types of directions and route finding. Furthermore, existing services allow users to tailor their own scenic routes and peruse tagged images of POIs online.
The authors differentiate their system by defining scenic driving as a “sightseeing activity that takes place in automobiles when people are traveling.” The aim therefore is to construct a scenic route that presents only geographical landmarks of interest that are visible from a moving vehicle.
The algorithmic calculations for mining and selecting tourist imagery, and the evaluation of roadside visibility produce an interesting discussion. The authors conclude the paper by identifying other critical challenges, such as multilane highways and streets containing traffic islands or barriers. They also identify the variations in attractiveness of scenery due to seasonal changes as another area for consideration.
This is interesting research that would be a useful and intuitive add-on for existing systems. Anyone working in GPS mapping may find it interesting. However, there is still some way to go with this project.
