Cloud computing assisted smart surveillance based safe transportation system to improve crime assistance on smart cities

• Ano de defesa: 2015
• Autor(a) principal: Camboim, Hugo Barros
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Brasil; UFRN; PROGRAMA DE PÓS-GRADUAÇÃO EM SISTEMAS E COMPUTAÇÃO
• Palavras-chave em Português: Segurança pública inteligente; Computação ubíqua; Processamento de vídeo; Computção em nuvem; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO
• Link de acesso: https://repositorio.ufrn.br/jspui/handle/123456789/22077

Smart video surveillance plays a key role in offering technological assistance in public safety scenarios, due to its potential to allow events and objects to be automatically detected in realtime. The Smart video surveillance technology has attracted special attention in the field of public safety, particularly in vehicular environments by the ability in assisting mission critical scenarios through real-time video processing based automatic detections (e.g., assaults, kidnaps, acts of violence etc.). Hence, it is possible to enable a tradeoff to be made between reactive and pro-active authorities when taking action, seeking to keep customized alert systems, carry out more efficient planning, and thus provide society with a better quality of life. Moreover, the association to Cloud Computing capabilities, is likely to provide advanced perspectives to smart surveillance applications, in terms to provisioning enhanced: (i) processing, to allow for deploying smart computing techniques for improved accuracy; (ii) storage, to adopt a wide variety of templates; (iii), and ubiquitous access, to keep connected everywhere and to any device. The investigation carried in this work focus in deploying a smart surveillance driven cloud-enabled approach to allow for event-based mobile applications in the public safety field. The objective is to allow optimized mobile applications, by which exploiting event-based scheme to provide to low resource consumption (energy, CPU and network usage) perspectives in comparison to the classic reactive approach. The evaluation is carried out in a set of experiments over a real testbed, which demonstrates that the proposal outperforms the classical mobile computing scheme.