Processamento embarcado de imagens com baixos níveis de contraste e iluminação para realce de objetos ocultos

Abstract

In this work, we present a study on image enhancement with very low levels of contrast and lighting, to help the recognition of hidden objects. The application is aimed in equipments with limited computational power, such as RaspBerry PI. The proposed method uses stochastic resonance combined with histogram correction that provides a final image with the expected quality. In order to carry out this study, researches were carried out in several papers in which the stochastic resonance technique was used either individually or sumed with other methods, as described in the related works review section. The results obtained were compared with techniques using artificial intelligence, and the results were quite interesting. The set of images used in this research consists of own images and public images used to test and train the methods of artificial intelligence.