Object Detection and Tracking for ASV

• Autor(a) principal: Patel, Dimpi Rajubhai
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/16747
• topic: ASV; ROAZ-II; Stabilization; Object detection

In this thesis automatic Object Detection system is presented. Object Detection is performed by different algorithms. As reading many literature we have observed that detecting objects in particular video sequence or by any surveillance cameras is a really challenging task in computer vision application because in sea the atmosphere affects a lot in the detection. Therefore we felt that there can be a wide range of possibilities are open in relation to detection. In order to improve the object detection, we developed image stabilization software on top of the image acquisition. First image stabilization has been performed over the raw data of ROAZ II. After achieving stabled video or images, object detection algorithm is performed using color based segmentation. Field tests have been performed with a data set from the ROAZ-II and during it shows the effectiveness of the approach. And system is able to achieve object detection in video or images with high accuracy.