A Semi-Automated Method for Monitoring of Roads with Traffic from High Resolution Satellite Image

Abstract

Efficient traffic management is critical in today’s overcrowded urban environments. Thanks to modern satellite imagery, high resolution images of current traffic conditions at regular intervals are easily available. In order to be able to respond to changes in traffic conditions quickly, it is critical to have an efficient and accurate system for traffic analysis of the static imagery available from these high resolution satellites. Two sub-problems are involved : detection of roads and detection of vehicles. A variety of algorithms are already available for solving these problems individually. For road tracking, we have median filers, color box segmentation, morphological processing, etc. There are algorithms like Bayesian background transformation, Blob extraction, and other vision based algorithms for vehicle detection. The goal of this paper is to automate tracking of roads and vehicles together from high resolution satellite images at any time statically, that is, without motion detection and leaving areas which are not roads like parking lot, settlements etc. Doing so, key points can be formalized in urban planning like finding other routes, human health factors, construction in endangered areas etc. By achieving the targeted objective and comparison with different algorithms, many problem statements can be defined and can be worked upon towards developing a solution.