Automating road junction identification using Crowdsourcing and Machine Learning on GPS transformed data

Identifying road junctions is of great importance for a number of applications that utilize electronic maps, like navigation systems. State of the art research on this area utilizes aerial images (usually captured by satellites), on which different image processing techniques are applied for automatically identifying road junctions. In this work, we propose a radical new approach to solve this problem. Instead of images, we propose an approach that relies on transformed Global Positioning System (GPS) data collected and analyzed using big data techniques. In particular, we apply machine learning on Crowdsource collected and annotated GPS data for automatically identifying junctions. Results suggest that the proposed technique is extremely effective. Furthermore, it is shown that it can be effective for solving the limitations that current approaches have.

Recommended citation: Your Name, You. (2024). "Paper Title Number 3." GitHub Journal of Bugs. 1(3).
Download Paper