Sistema de processamento de big data aplicado à monitoração de automóveis em tempo real

Abstract

The high number of vehicle thefts together with the low number of cars that have some tracking solution make it essential to create vehicle identification and recognition solutions. Computer vision systems can be allies in monitoring public areas like streets and highways. However, vehicle analytics is a challenging task due to many factors, including diversity of vehicles, brand and model features, color, direction of movement, image position and lighting changes. In addition, the computational cost of detection systems is high. This course conclusion work describes the strategy adopted for the design and development of a system to identify vehicles and index them through license plate, brand and color. This system will be integrated with a Big Data processing system, allowing the information extracted to be made available for queries based on these characteristics.