INFORMATION ACQUISITION FOR CITY TRAFFIC MODELS BASED ON IMAGE ANALYSIS

摘要

Fast development of cities and increasing number of cars cause that metropolises need to improve their monitoring system. This paper presents the system which gathers information about traffic and help to monitor and model it. It is not as expensive as current solutions, because it is built on existing infrastructure and low-cost devices. The basic part of this system is a smart phone, on which application is implemented. It contains capturing device and processing unit, as a result it returns information about traffic. Information among other things includes time of the appearance, type of a car and its speed. Image analysis (processing) consist of pre-processing (geometry reconstruction and denoising), classification (geometrical methods), object detection and tracking algorithm, which based mostly on calculating gradient. Thanks to this solution, monitoring system can be easily implemented on the places, where it was not possible because of high cost and problem with creating infrastructure.